Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
-
Upload
cristina-icanu -
Category
Documents
-
view
218 -
download
0
Transcript of Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 1/375
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 2/375
Generating Social Stratification
Toward a N ew Research Ag end a
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 3/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 4/375
Social nequality Series
Marta T ienda and David B. Crusky Series Editors
Generating Social Stratification: Toward a
New Research Agenda edited by Alan C. Kerckhoff
The New Role of Wom en: Family Formation
in Modem Societies edited by Hans-Peter Blossfeld
Social
Stratification: Class Race and
Gender
in Sociological Perspective edited by David B. Grusky
Careers and Creativity: Social Forces in the Arts
Harrison C. White
Persistent Inequality:
Changing Educational Attainment in Thirteen Countries
edited by Yossi Shavit and Hans-Peter Blossfeld
The Arab M inority in Israel s Economy: Patterns of
Ethnic Inequality Noah Lewin-Epstein and Moshe Semyonov
Equality and Achievement in Education
James S. Coleman
Ethnicity and the New Family Economy:
Living Arrangements and Intergenerational Financial Flows
edited by Frances K. Goldscheider and Calvin Goldscheider
FORTHCOMING
Social Differentiation
and
Social Inequality: Essays
in Honor of John
Pock
edited by James Baron,
David B. Grusky, and Donald Treiman
Education and Social Class in Comparative Perspective
edited by Robert Erikson and Jan O. Jonsson
Inequality and Aging John Henretta and Angela O Ra nd
Prejudice
or
Productivity:
Ethnicity
Languages
and
Discrimination
in
Labor
Markets M.D.R, Evans
Between Two Worlds: Southeast Asian Refugee
Youth in America Ruben G. Rumbaut and Kenji Ima
Children Schools and Inequality
Doris R. Entwisle and Karl Len Alexander
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 5/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 6/375
Generating
Social Stratification:
Tow ard a N ew Research A ge nd a
EDITED BY
Alan C Kerckhoff
Duke U niversity
~ r A Member of the Perseus Books Group
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 7/375
ocial nequality eries
All rights reserved. Printed in the United States of America. No part of this publication may
be reproduced or transmitted in any form or by any means, electronic or mechanical, including
photocopy, recording, or any information storage and retrieval system, without permission in
writing from the publisher.
Copyright © 2000 by Westview Press, A Member of the Perseus Books Group
Published in 2000 in the United States of America by Westview Press, 5500 Central Avenue,
Boulder, Co lorado 80301-2877, and in the U nited K ingdom by Westview Press, 12 Hid s Copse
Road, Cu mn or Hill, Oxford OX2 9JJ
Library of Congress Cataloging-in-Pubfication Data
Generating social stratification : toward a new research agenda /
edited by Alan C. Kerckhoff
p. cm.—(Social inequality series)
Includes bibliographical references.
ISBN 0-8133-8967-4 (he) ISBN 0-8133-6796-4 (pbk) .
1. Students— United States—Social condition s. 2. Academic
achievemen t—U nited States. 3. Social mobility—U nited States.
I. Kerckhoff, Alan C. II. Series.
LC205.G46 1996
371.8*0973—dc20 95-45691
CIP
The paper used in this publication meets the requirements of the American National Standard for
Permanence of Paper for Printed Library Materials Z39.48-1984.
10 9 8 7 6 5 4 3 2 1
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 8/375
Contents
Preface xi
P RT ONE
Conceptualizing Careers and Stratification Processes
Introduc tion to Part One 1
1 Structuration and Individualization: The Life Cou rse
as a Co ntinu ous Multilevel Process
ngela
M . O Rand,
Duke University 3
2 Social Psychological Aspects of Achievement
Jeylan T Mortimer, University of Minnesota 17
3 Building Concep tual and Empirical Bridges betw een
Studies of Educational and Labor Force Careers
Alan C. Kerckhoff,
Duke University
37
P RT TWO
Educational Contexts and
Processes
Introduction to Part Two 57
4 Educational Stratification and Ind ividu al Careers
Adam G amoran
University
of Wisconsin
Madison 59
5 Educational Tracking du ring the Early Years:
First Grade Placemen ts and M iddle School Constraints
Karl
L .
lexander and Doris
R.
Entwisle
The Johns Hopkins Univerisity
75
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 9/375
lll
6 Peer Social Netw orks and Adolescent Career
Development Charles E. idwell
and Stephen Plank The University of Chicago;
Chandra
Midler
The University
of Texas Austin
107
7 School Choice and Co m m unity Segregation:
Findings from Scotland /. Douglas Willms
University
of New
runswick and University
of Edinburgh
133
8 Educational Processes and School Reform
Maureen T
Hallinan
University
of
Notre Dame 153
P RT THREE
Education and Labor Force Linkages
Introduction to Part Three 171
9 Educational Credentials and the Labor M arket:
An Inter-Industry Comparison
W.P. Bridges
University
of
Illinois
a t
Chicago
173
10 Education Earnings Gain and Earnings Loss in
Loosely and Tightly Structured Labor Markets:
A Com parison between the United States and G ermany
Thomas A .
DiPrete
Duke
U niversity;
Patricia A. McManus Indiana University 201
11 Education and Credentialing Systems Labor Market
Structure and the Work of Allied H ealth Occupations
Robert
Althauser
Indiana University;
Toby Appel Yale University 223
12 Creating Cap italists: The Social Origins of
Entrepreneurship in Post-Communist Poland
arbara
Heyns New
York University 257
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 10/375
X
P RT FOUR
Social System Contexts
Introduction to Part Four 291
13 The Politics of Mobility
Michael Hout University of
C alifornia
erkeley 293
14 Stratification and A ttainm ent in a Large Japanese Firm
Seymour
Spilerman
Columbia University;
Hiroshi Ishida University
of Tokyo
317
15 Ch anging Contexts of Careers: Trends in
Labor Market Structures and Some Im plications
for Labor Force Outcom es Arne L .
Kalleberg
University
of North
Carolina
at
Chapel
Hill
343
About the ook and Editor
359
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 11/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 12/375
Preface
Ganzeboom , Treiman and Ultee 1991) refer to three generations of social
stratification research . In the first generation, research wa s concerned w ith
com paring tables that cross-classified categories of the occupational posi
tions of fathers an d sons as they varied across societies and over time. They
see the second gen eration as having begun w ith Blau and D uncan's (1967)
introduc tion of a basic m od el of status attainm ent. That model shifted the
focus from father-son com parisons to stages in a process by which sons are
distributed in a continuous status hierarchy. The third, most recent,
generation of stratification research, they observe, has returned to the
earlier focus on tables of parent-child locations in occupational or social
class categories, althou gh the current analyses are much m ore intricate d ue
to mo re sophisticated theories, new analytic mod els, and mu ch better d ata.
This volume could
be
viewed a s
a
continuation of wh at Ganzebo om et al.
call the second generation , bu t I prefer
to see
it
as
part of an emerging fourth
generation of stratification research. It clearly ha s many of the features of
the second generation, but the kind of research reflected in this volume
mo ves us well bey ond the second. An d, as with the com parison of the first
and third generations, the advance has been due to improvements at the
level of theory, analytic techniques, and improved data sources.
The difference be tween the first and third generations, on the one han d,
and the second and fourth, on the other, is the focus of the latter two on
stratification processes. Rather than modeling pattern s of interg enera tional
or career m obility, the goal is to identify and take account of the processes
by which a birth cohort
is
distributed in the stratification system, given their
positions of origin in that system. It is increasingly a ppare nt that this is an
awesome task, and we are far from reaching that goal. Yet, there have been
imp ortant gains durin g the past two decades, many of which are reflected
in the chap ters of this book.
Ever since Blau and Duncan's (1967) introduction of a path model
conceptualization, there has been a tendency to look at the overall stratifi
cation process as having three cum ulative stages: the process by which
social origins affect education al attainm ents; the process by w hich origins
an d ed ucation al a ttainm ents affect the level of labor force en try; the process
by which origins, educational attainments, and first jobs affect workers'
moves from their initial pos itions to later positions in the labor force. The
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 13/375
ll
Blau-D uncan three-s tage m ode l served to descr ibe the re la t ionships am on g
origin , edu catio nal at tain m ent, f irst job, an d later job, bu t i t did no t tell u s
very much about how those re la t ionships come about .
A t tem pts w ere soon ma de to expla in the re la tive ly s t rong re la t ionship
betw een socia l or igins and educa t ional a t ta inm ents . The W iscons in mod el
of s tatu s at ta inm en t (Sewell et al . 1969) sh ow ed ho w social psyc holo gical
m ea su re s (am bition , influence from significant others), reflecting family
an d school proce sses, he lpe d to explain that relat ionsh ip. M uc h less w as
do ne to analyze the second a nd thi rd s tages of the a t ta inm ent process , a nd
som e w riters (e.g. , H ora n 1978) qu estion ed the adequ acy of a social psy ch o
logica l app roach to expla ining the s ta tus a t ta inm ent process . They arg ued
tha t the re are features of the social stru ctu re th at alter the free influe nce of
personal qual i t ies and interpersonal re la t ions on individuals ' achieve
ments .
Such cr i t iques served to direct at tention to s tructured contexts and
allocation processes. They also shifted the focus from the first stage to the
later s tages of the Blau-Dunc an m od el bec ause they we re largely c onc erned
w ith the s tru ctu red na tu re of the labor force. They arg ue d that the labor
force is far f rom an ope n op po rtun ity s tructu re, an d whic h jobs in div idu als
enter and move between are de termined by more than the ir personal
character is t ics .
Ind ust r ial sectors pr ov id e different op po rtu nit y s tructu res (Piore 1975).
Internal labor m arke ts , cons is t ing of f i rms ' job ladd ers and pro m otio n
regim es, favor those already in the f irm (A lthauser an d K alleberg
1981),
and
as those at the low er levels on the job ladd er mo ve u p, this creates a vaca ncy
ch ai n (Sorense n 1977). Since no t all jobs in a firm are on job lad de rs,
ho w ev er , som e op enin gs are close d to all w ho are not at the nex t low er
level on a job ladd er wh i le others are op en to outs id ers (Sorensen an d
T u m a 1981). The central po int of these analyses wa s that w or ke rs ' location s
in the highly structured labor force strongly affect their opportunities for
mob i l i ty du r ing the ir work careers .
Similar s tructural effects were identif ied in the educational at tainment
proc ess . In par t icu lar , s tud en ts ' locations in the abil ity grou ps or tracks in
schools were shown to have s ignif icant effects on both their aspirations
(Heyns 1974) and their achievements (Alexander , Cook and McDill 1978) .
Here too, i t was argued, the s tructure of opportunit ies within schools
affec ted b oth th e individ uals ' person al qual it ies and the outco me s av ai lable
to them . A nd , in m uc h the sam e w ay that internal labor m ark ets l imit access
to favored structural locations in the labor force, access to the favored
locations in the school s truc ture is also far f rom fully op en (Hall ina n an d
Sorensen 1983).
A s Baron (1994:390) has obse rved, ho w eve r: W e ha ve m an y mo re
coherent f indings now about how specif ic var iables affect mobili ty or
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 14/375
lll
attainment than we have coherent stories about why (and under what
circumstances) those results obtain. One of Baron 's concerns is that the
new struc turalism has been too narrow ly concerned with the structural
contexts within which social mobility occurs and has ignored the social
psychological processes that help explain the effects of those contexts.
As research has dem onstrated the interplay am ong person al qualities,
structured contexts, and outcomes, the conceptualization of stratification
processes has become m ore and m ore
complex.
It has become necessary to
take into account cum ulative processes that occur throug hou t the formative
years as well as du ring w ork ers ' careers in the labor force. There have been
increasing attem pts to take into account the interplay betw een institutional
arrang em ents and social psychological dynam ics.
As more and better da ta sources have become available and the facilities
and methods for multivariate analyses have become more widespread,
cross-societal com parisons have become feasible, and societal differences in
even the basic form of the status attainment m ode l are appa ren t (Treiman
and Yip 1989). There are notable societal differences in the specific pro
cesses involved, as demonstrated in several of the contributions to this
volume.
In part because of the increasingly complex conceptualizations and
observations and the difficulty of encompassing the full range of relevant
factors in any single analysis, there has also evolved a division of labor in the
stu dy of stratification processes. One sharp division is between those who
focus on stratification p rocesses tha t occur early in life, especially in educa
tional institutions, and those who focus on processes within the labor force.
It was my concern about that division that led me to organize the
conference at wh ich the papers in this volume were presen ted. The goal of
the conference w as to bring together scholars who have specialized in the
stud y of one or the o ther of these two parts of the stratification process in
order to increase mu tual interest in each other's w ork and to provide a basis
for greater integ ration of the two parts both conceptually and empirically.
We wanted to resensitize specialized scholars, including ourse lves, to the
nature of the entire stratification process and to how each of our more
focused efforts contributes to an overall und ers tandin g of the full process.
We hope that this volume will contribute to a broa der recognition and
apprec iation of
the
challenge of developing a mo re coherent und erstan ding
of stratification processes. It is the attempt to develop such a coherent
understanding, firmly founded on a focus on institutional arrangements
and the associated social psychological processes, that leads m e
to
view this
book as a contribution to an emerging fourth generation of social stratifica
tion resea rch. believe that the papers in this volum e do m ake a con tribu
tion to such an un derstand ing and point the way to further clarification. I
anticipate a con tinuing effort to build on this initial contribution.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 15/375
x v
The conference w as held d uring a three-day period in April 1994. It
began with a session at the annual meeting of the Southern Sociological
Society in
Raleigh
N.C. on April 9th and continued in a series of sessions on
the Duke University cam pus on April 10th and 11th. The conference was
m ade possible and the publication of
this
book w as significantly facilitated
throu gh generous suppo rt from The Ho ward E. Jensen Fund and add i
tional grants were received from the Provost and th Dean of Faculty of Arts
and Sciences of Duke University. All of us who pa rticipated in this joint
effort are very grateful for this financial support.
Several people made im portant contributions along the way. Judith
Dillon helped immensely in organizing the conference. Elizabeth Glennie
and Lorraine Bell participated in and handled many of the details of the
conference. Lisa Shander took on the challenging task of collating the
individu al papers into a well-organizedm anusc ript. The entire enterprise
depend ed heavily on their assistance and I am pleased to offer them my
wa rm thank s. Finally I very m uch appreciate
th
interest
in
the book show n
by Marta Tienda and David Grusk y the editors of this series and the
patience of Dean Birkenkamp the W estview editor in working through the
complexities of bringing the book into existence.
Alan C. Kerckhoff
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 16/375
X
eferences
Alexander, Karl L., M artha A, Cook, and Edward L. McDill. 1978. Curriculu m
Tracking and Educational Stratification.
American Sociological
Review43:47-66.
Althauser, Robert P. and A me L. Kalleberg. 1981. Firms, Occupations and the
Structure of Labor Markets:
A
Conceptual Analysis and Research A gend a. Pp.
119 49 in
Sociological Perspectives on Labor Markets
edited by Ivar Berg. New York:
Academic Press.
Baron, James N. 1994. Reflections on Recent Generations of Mobility Research.
Pp. 384 93 in
Social Stratification:
Class Race
and Gender in Sociological Perspective
edited by David B. Grusky . Boulder, CO: W estview Press.
Blau, Peter M. and Otis Dudley D uncan. 1967.
The American Occupational
Structure.
New York: Wiley.
Ga nzeb oom , Ha rry B. G., Don ald J. Treiman an d W oot C. Ultee. 1991.
Intergenerational Class Mobility in Comparative Perspective.
Research
in
Social Stratification and Mobility
9:277-302.
Hallinan, Maureen T. and Aage
B. Sorensen.
1983. The Formation and Stability of
Instructional Groups.
American Sociological Review 48:838-
51.
Heyns, Barbara. 1974. Social Selection and Stratification W ithin Schools. Ameri-
can Journal o f Sociology 79:1434-51.
Horan , Patrick M. 1978. Is Status Attainm ent Research Atheoretical?
American
Sociological Review 43:534-41.
Piore, Michael. 1975. No tes for a Theory of Labor M arketStratification . Pp. 125-50
in
Labor Market
Segmentation edited by Richard C. Edwa rds, Michael Rich, and
David M. Gordon. Lexington,
KY:
Heath.
Sewell, William H., Archibald O. Haller and A lejandro Portes. 1969. The Educa-
tional and Early Occupational Attainme nt Process.
American
Sociological Review
34:82-92
Sorensen, Aage
B.
1977. The Structure of Inequality and the Process of Attain
ment.
American Sociological Review
42:965-78.
Sorensen, Aage B. and N ancy B. Tum a. 1981. Labor Market Structures and Job
Mobility.
Research in Social Stratification and Mobility
1:67 94.
Treiman, Donald T
v
and Kam-Bor Yip. 1989. Educa tional and Occupa tional
Attainment in 21 Countries. Pp . 373-94 in Melvin L. Kohn (ed.),
Cross National
Research in Sociology. New bury Park, CA: Sage.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 17/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 18/375
PART ONE
Conceptual izing Careers
and Stratification Processes
The emerging conceptualizat ion of the processes by which societal
s trat if icat ion is generated includes a considerat ion of factors at mult iple
levels of analysis ranging from the dynamics of individual functioning to
the na tu re of the society 's econo m ic and po li tical system . The challenge w e
face is to integrate our knowledge of these mult iple levels in ways that
clarify the nature of s trat if icat ion generat ing processes . The three papers
inc luded in Par t
sug gest som e of the features of the bro ad con ceptu alizat ion
w e need . In this w ay , they help to set the s tage for the other p ap er s th at
follow.
In the firs t p ap er in Part I Ang ela O 'R and em pha s izes the imp ortanc e of
this m ult i leve l con ceptu alizat ion . She rem ind s us that the l ife cou rse is an
on-going process whose shape is dependent on mul t ip le sources of inf lu
ence. She succinctly reviews bodies of literature dealing with the full l ife
course , showing how ins t i tu t ional and individual fac tors he lp to produce
the obs erve d trajectories. She sho w s ho w wh at she cal ls the trad it ion al
ins t i tu t ional ap pro ach es d if fer f rom re la t ional ap pro ach es and ho w
they com plem ent each other in contr ibut ing to an un der s tan din g of the ro le
of social structure in shaping the life course.
The second paper , by Jeylan Mort imer , concentra tes on the soc ia l psy
chology of achievement, emphasizing the s ignificance of social izat ion
processes both within the family and outs ide i t . Mortimer r ightly argues
that i t is not a quest ion of choosing between a s tructural and a social
psychological approach to achievement; i t is essential that we see the
con nectio ns betw ee n the two . She also po ints out that , of the tw o bo die s of
l i terature dea ling w ith educ ation a nd labo r force processes , the l i terature on
the labor force ha s be en less con cern ed w ith the role of social psyc holo gica l
processe s . She spec ulates about how educ ation al an d labor force exp eri
ences can a l te r a t t i tudes an d va lues and asks us to th ink m ore de eply an d
clearly about ho w s t ruc tures and soc ial psychologica l processes a re in terre
lated.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 19/375
In the third pa per in Part I, Alan Kerckhoff directs our a ttention to the
struc tured contexts of individual careers, devo ting much of the discussion
to an analysis of impo rtant differences in those contexts among indu strial
societies. The effects of both educational and labor force contexts are
discussed, but perh aps the most impo rtant issue dealt w ith is the societal
variation in the natu re of the linkages between those two contexts. Kerckhoff
uses the identified inter-societal differences to generate some tentative
hyp otheses that deal with social psychological effects on ind ividuals s well
as overa ll societal patte rns of social stratification.
Toge ther, these three papers prov ide an indication of the bread th of the
research agenda to wh ose developm ent this volum e is dedica ted. Each of
them makes reference to the other papers in this collection, but they also
po int up additiona l kinds of investigation that are, or need to become, part
of the overall effort. The papers in Parts II, III and IV po int to the
com plexities of societal stratifying processes. They bring into focus the
general obse rvations O Rand makes about the life course being
multilevel
process. They provide nu m erous examples of the need to take both social
structural and social psychological processes into account, as Mortimer
sug ges ts. An d, they dem ons trate the imp ortanc e of the kind s of
inter-organizational linkages dealt with by Kerckhoff as well as the global
characteristics of the societies studied.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 20/375
Structuration and Individualization:
The Life Course as a C on tinuou s,
Multilevel Process
ngela
M.
O Rand
Duke University
Sociologists hav e long recogn ized the prom ise of cross-level analysis for
l inking indiv idua l beh avio rs to social s truc tures an d for d is t ingu ishing the
sepa ra te processes opera t ing betw een
levels.
H ow eve r , unt i l recently m eth
odological obstacles have hindered the successful achievement of this
promise. Theoretical exposit ions on the transposit ion between levels of
an alys is ex ten d back to the early fifties, inclu din g a pa rticu larly insigh tful
and prophetic discussion by C. Wright Mills (1953) who anticipated the
pivo ta l ro le of meth odolo gica l as oppo sed to conceptual inn ovat io ns for the
ult imate resolution of these issues. The specif ication of the properties of
collectivities at successive levels of analysis (e.g. Coleman, 1970) and the
separation of structural from individual effects (e.g. Blau, 1960) are among
the ear l iest and most persis tent conceptual concerns that are now arr iving
at som e m etho do log ical so lution (e.g. H ub er , 1991; DiPrete an d Forr is tal ,
1994).
Social tran sitio ns over the life co urse pro vi de strate gic subject ma tter for
cross- level analyses . The movement of individuals wi thin and between
insti tutional contexts br ings into focus how lives are shaped at social
interfaces and , in tur n, ho w insti tution s ma y them selves be transform ed in
response to the social f r ict ion exer ted by demographic processes, i .e . by
agg regate pa t t erns of individ ual t rans i t ions (O Ran d, 1995). A s su ch,
cross- level ana lysis reveals wh at An tho ny G idd en s refers to assf ructuration
or ho w social s t ruc tures are bo th cons t i tu ted by h u m an agency an d a t the
same t ime the medium of this consti tution (Giddens, 1976:121) .
Studies of l ife-course transit ions—from the transit ion to adulthood,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 21/375
4
through midlife transit ions associated with employment or labor force
shif ts and family change, to the transit ion to ret irement—are revealing
patterned var iat ions in their trajector ies and temporal organizations that
em erge f rom the in terac t ion betwe en ins t i tu t ional contexts and ind ividu al
bio gra ph ies (M ayer an d Tu m a, 1990; Elder an d O 'Ra nd , 1994). These
pa th w ay s of ind ivid ual trajector ies indica te w ha t migh t be referred to as the
essential tension betw een the s tructu ration a nd the ind ividu aliza tion of the
life course across insti tutional contexts . As such, the continuous, cumula
tive proc ess of the life cours e org ani zed by m ultile ve l forces is a strate gically
appropriate domain for cross- level analysis .
In the pres en t essay, recent exa min ation s of life course pro cesses will be
reviewed in light of their implications for the study of both the variability
of ind ivid ua l trajectories and the insti tution al pattern ing of var iabil i ty. Tw o
them es in these s tudies w ill be con sidere d. The firs t is ind ivid ualiz ation , o r
w ha t is referred to in aging research as the thesis of hete roge neity , an idea
suppor ted pr inc ipal ly by longi tudinal demographic research on popula
tions in advanced western societies (particularly the U. S.) . Second, is the
specif ication of social s tructure. Here there are al ternative approaches to
s t ructura t ion: chief am on g these are w ha tm ig h tb e refer red to as t radi t ional
ins t i tu t ional exam inat ions and re la t ional appro ache s . Both hav e d eve lope d
rapid ly in the ne w s t ructura l ism benefi t ting f rom m ethodo logica l innov a
t ions in mul t i level model ing and network analys is and f rom growing
dat aba ses that sp an social contexts . Each m ak es a dis t inctive contr ib utio n to
the analys is of s truc ture.
Ind ivid ua lizatio n Ov er the Life ourse
The focus on transit ions am on g life ph ases and the s truc tural interfaces
that shap e these t rans i t ions is reveal ing the dem og raph ic a nd ins t i tu t ional
mech anisms tha t o rgan ize lives. It is also unveil ing the considerab le he tero
ge ne ity of life trajectories. We ha ve lea rne d, for ex am pl e, tha t the so-called
transition to dulthood in the Un ited States is not a cr isp, sequ ence of s tatu ses
fol lowing s t rong age or t iming norm s (Hog an and A stone , 1985; Rindfuss,
1991).
Rather i t i s a dem ographica l ly den se yet tempo ral ly he te rog ene ou s
order ing of multiple transit ions ( including leaving school, leaving home,
star t ing work, forming households, becoming a parent) that is dr iven less
by age norms per se than by ins t itu t ional ized l inkages or pa th w ay s across
educat ion, work, and family domains tha t d i f ferent ia te among subgroups
of the adolescent population.
John M od ell (1989) arg ue s that the trend tow ard s the no rm ativ e loosen
ing of this transit ion has em erg ed steadily over this cen tury. His social
his tory of the transit ion from yo uth to ad ult ho od be tw een 1920 and 1975—
aptly titled Into One s O wn —documents the exte nsio n of ed uc atio n later
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 22/375
into the life course, changes in marital and parental timing and their
determ inants, changes in the transition
to
stable em ploym ent, and changes
in the propensity for marriage and divorce and in their covariates (espe
cially pare ntho od) , among other trends. O verall, he observes that m en's and
w om en 's life-course trajectories in family and w ork participation pa tterns
are becom ing increasingly similar and tha t
the
life paths betw een black an d
wh ite popula tions in which earlier inequalities app ear to be exacerbated
over time by structural opportunities and dem ographic p atterns are becom
ing more dissimilar.
The master trend, according to M odell, has been in the liberating of one
transition from the next—the weakening...of determinate sequences and
interva ls by the injection of increasing volition into the youthful life
course (Modell, 1989:332-333). The ever-grow ing availability of economic
and cultural resources provided to the young by family, govern mental, and
market structures is argued by Modell to be the principal source of this
intercohort trend. But
his
argum ent accounts less rigorously for w ithin-cohort
variations and divergence among class, race and gender categories. The
stratified bases of volition and the primary institutional contexts that
define the options available to segments of succeeding cohorts of U. S.
adolescents over this century are less clearly de lineated by him .
Studies reported in this volum e w ould suggest mat the increased hetero
geneity in the transition to adu lthood in the
U.S.
may be in part accounted
for by the jointly influential phenomena of (1) expansion of educational
opportunities; (2) the stratification of the educational institution along
class-related ability g roup ings or tracks as well as across priva te and public
school contexts; an d 3) the weak coupling of educational and em ploym ent
institutions. Paradox ically, school-to-job linkages have grow n increasingly
ambiguous as m ore and more adolescents achieve secondary school educa
tions.
All in all, the im pact of schooling on early life-course trajectories is a
model of cross-level influences on divergent life pathways Kerckhoff,
1993).
i life—a
life ph ase once typified as dem ograp hically sparse— appears
actually
to
be more and more punc tuated by status transitions that increase
the heterogen eity of life trajectories. And as we are finding for the transition
to ad ultho od, these transitions have diverse tem poral features; their timing,
ordering, density, even reversibility yield multiple trajectories. Family
formation and dissolution, movem ents into and out of jobs, and changing
intergenerational dependen cies perv ade the adu lt life course. These hetero
geneities, like those evident in the transition to adulthood, appear to be
products of the interplay among demographic and structural processes
channeling individuals. And, they are prod ucing divergent socioeconomic
trajectories anchored in early opportunities but deflected by adult transi
tions related to wo rk, family, health an d other contingencies of adu lthood .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 23/375
6
Even the retirement transition is becom ing more demog raphica l ly var ied
as i t s boundar ies are be ing pushed backward and forward by revers ible
transit ion s am on g wo rk, family, leisure an d health s tatu ses over a m ore an d
m ore ex tend ed per io d of life (Dannefer ,
1988).
Ret i rement is no t a na r row ly
demarcated, absorbing, and determinate sequence of events def ined by
straightforward exits f rom work. We are identifying multiple pathways in
an d ou t of w ork that follow different sched ules in later yea rs . Em plo ym en t
in pos t-career or br idg e jobs, par t ia l and intermit tent w ork, and disabi li ty
are am on g several al ternativ e rou tes to ret irem ent that can be ad d ed to the
t r ad i tiona l pens ion and hea l th pa th w ays s tud ied fo r 80 m an y yea r s am on g
older ma le w orkers . These m ul t ip lying p at terns ema nate f rom the in terplay
of bi og rap hy an d social chan ge (H enretta , 1992; O 'Ran d, 1995; Elder an d
O'Rand, 1995).
The Thesis of eterogeneity
De m og raph ers , epidemiolog is ts , social h is tor ians and sociologis ts h av e
observed patterns of increasing differentiat ion within aging cohorts f rom
bir th to death (Dannefer , 1987,1988,1991). These patterns are constructed
by sequen tial ly conting ent s tatus transit ions or life pa th w ay s in w hich later
transit io ns are con straine d by earl ier
ones .
O ve r t ime, the gen eral pa tte rn of
differentiat ion produces diverse outcomes of interest , including individu
a l iza t ion and inequal i ty . The most s t r ik ing pat terns of individual iza t ion
have been found in the s tudy of health and i l lness in aging populations
(M add ox , 1987). He alth-related exp eriences and beha viors interact w ith
other l ife transit ions with cumulative and relat ively highly individualized
outcomes in old age. Intertwined life trajectories of health, work, family,
leisure ref lect highly individualized exposures to opportunit ies and r isks
related to health and functioning in later life.
Similar ly, older populations exhibit greater social and economic in
equ ali t ies tha n yo un ge r age gro up s du e to the cu m ula tive effects of differ
ent ia l oppor tuni t ies and achievements over t ime. For example , conven
tional measures of income inequali ty across age groups (specif ically gini
coeff icients) repeatedly reveal higher levels of dispersion among older
gro up s (Crys ta l and Shea , 1990). The cum ulat ive a dv ant age hyp oth es is ,
initially proposed by Robert Merton (1968) to explain the high levels of
ineq uali ty in pro du ctivity and recognit ion am on g scientif ic careers , is no w
more widely applied to s tudy aging or l ife-course processes of social
inequal i ty l inking individ ual behaviors w i th s t ruc tura l ph en om en a re la ted
to sys tems of advantage and disadvantage tha t cont inuous ly s t ra t i fy co
ho rts ove r t ime (see Dan nefer , 1987,1991).
Social psyc holog ists have observ ed across multip le s tud ies that s tabil i ty
coefficients of personal characteristics such as attitudes, performance lev
els, and patterns of social relat ions increase with age (Elder and O'Rand,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 24/375
1995 for review; Alwin, 1993). Here it is observed that patterns of
self-definition, actualization or legitimation are sequentially reinforced or
otherwise reestablished as individuals m ove across oppo rtunity contexts.
Short du ratio ns between transitions appear to increase stability coefficients;
longer durations seem to attenuate them.
The social psychology of achievem ent appears to reflect similar persis
tence across contexts. At a structura l level, achievem ents in earlier struc
tural contexts regu late the access to new contexts of achievement. Accord
ingly, the interfaces betw een institutional env ironm ents, as those be tween
schools and jobs or between jobs, constrain individual transitions and
allocate individuals selectively between contexts Kerckhoff, 1993; see
Kalleberg s contribution to this volume).
At the individual level, earlier achievements enhance self-selection
processes via mechanism s related
to
self-esteem, iden tity form ations, social
competency and reinforced achievement m otivation. Thus, structural allo
cation and self-selection processes enhance intraindividual stability and
interindiv idua l heterogeneity over time (Alwin, 1993). Jeylan M ortim er s
essay in this volum e aptly depicts these processes in the social psychology
of achievem ent.
The growing he terogeneity of life pa thw ays following different transi
tion sequences and timing patterns is shedd ing m ore light on the complex
interplay of chang ing institutional arrangem ents and chang ing lives. Even
the idea of the triangularization or
the
tripartitioning of life into pre-work,
work, and non -wo rk phases correlated with age is losing its usefulness for
conceptualizing the social organization of the life course (Wilensky, 1981;
Kohli, 1988; Riley, 1993). Lives move along asynchronous tracks which
multiply with age, especially following the transition to adulthood, and
appear to produce a pattern of individualization. The synchronization of
educational, family, work, leisure, health and other transitions
is the
central
generating mechanism of the life course.
The institutional and contextual mechanism s for these changes after the
transition from secondary education are anchored in the workplace, the
family, and in the wider cultural and political-economic environments in
which these prox imate life contexts are em bedded (Mayer and Schoepflin,
1989).
And while we are learning that educational transitions may follow
the one from secondary education in adolescence in late-20th century
societies, the strong differentiating effects of this transition for later ones
continues to attract our attention. This volume
is
devoted to identifying the
structural mechanisms that operate to organize lives at the transition to
adulthood.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 25/375
Structuration Tw o App roaches to Op portu nity Structure
How does th is cumulat ive process l inking sys temic and biographica l
components take place? Is i t
a
s imple accumulat ion ( adding up ) of
individual s tatuses in succeeding contexts over t ime? Probably not. I f i t
were, then there would be less heterogeneity with age than we observe.
Insti tutional s tructures in modem societies shape the l ife course via indi
viduals ' s imul taneous par t ic ipa t ion in segmented and co-occur r ing role
schedules over t ime. They also provide the condit ions for the progressive
defin ition, a ctua lizatio n a nd legitim ation of the self (Koh li, 1986). Th us , the
seemingly contradic tory phenomena of the s t ruc tura t ion and individual
ization of ad ult l ives are o bserv ed.
The cum ulat ive l ife-course s bes t exam ined as the in terplay a m on g three
processe s: selection, op po rtun ity s truc ture, and t ime (see Rosenfeld, 1992,
for a
discussion of career processes as generally comprised of these ele
m e n t s ) .
Selection effects include the cum ulativ e biog raph ical con dit io nin g
that individuals br ing with them to succeeding contexts . Oppor tuni ty
s t ructures inc lude s t ruc tura l ly var ied and proximate t rans i t ion contexts
such as schools , jobs, and marr iages
nd
their wider insti tutional and
histor ical en viro nm en ts suc h as s tate policies relating to schools an d their
organiza t ion, em ploy m ent prac tices that t ie jobs and job rewa rds to creden
t ials, cit izenship obligations and enti t lements and other poli t ical-economic
cond i t ions . Cross-nat ional s tudies repeatedly d ocu me nt the imp or tan ce of
state policies related to education, labor, and social welfare for the level of
va r iab i l i ty o r inequa l i ty in ind iv idua l we l l -be ing and ach ievement
(Bu rkhau ser , Du nca n an d H aus er , 1994; Pam pel, 1993; H o ut in this vol
u m e ) .
Wars, economic cycles or transformations, and natural disasters can
also influence the distribution of resources and options at critical transition
periods in l ives and producing longlasting effects on achievement (Elder
and O'Rand, 1995) .
In add it ion , t ime acts as an ind ep en de nt force on the org aniz ation of the
life cou rse. The t iming and ord er ing of transit ion s and d ura tio ns in succeed
ing 's tates ' or contexts fur ther organize the cumulation process . As such,
cross- transit ion l inkag es com prise life pa th w ay s (Hog an and A ston e, 1986)
with cont inuous and discont inuous fea tures s temming f rom this in terac
tion and the temporal characteristics that further differentiate them.
Traditional Institutional Approaches
Cross- level analyses ha ve been spu rred on in the pas t tw o deca de s by the
introd uctio n an d elabo ration of the so-called ne w stru ctu ralism to s trat i
fication and social mobility research
Kerckhoff,
1994). The status attain
ment model of stratif ication, and its social psychological elaboration in the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 26/375
W iscons in m od el , dom inated research on achievem ent and inequal i ty
from the la te 1960's to the early 1980's. This ex em pla r ex am ine d th e family
or ig ins - to-educat ion- to-occupat ion d is t r ibut ions of indiv idua ls wi tho ut
expl ic i t t rea tm ent of the nor m at ive s t ruc ture s in wh ich they were em bed
de d (Kerckhoff, 1976). Co nse qu ently , the exp lanato ry fram ew ork ha d an
individual is t ic b ias tha t could no t account for subg rou p va r ia t ions (such as
those associated with race, gender and industria l location) in at ta inment.
The s t ruc tura l ism in t roduced ins t i tu t iona l covar ia tes of indiv idual d is t r i
bu tio ns that a t t em pte d to acco unt for these differences .
Normative domains such as educational systems (Kerckhoff, 1994),
industries and labor markets (see Kalleberg and Bridges in this volume),
and organ iza t iona l op po rtun i ty regim es (Law rence , 1984; Sorensen , 1986;
Sp ilerm an, 1977, 1986; an d Kalleberg in this volu m e) w ere am on g the
prin cip al inst i tut ion al contexts inco rpo rated ini tia l ly in s tratif icat ion analy
ses as ind iv id ua l a t t r ibutes . They w ere ass igned va lues as g lobal cova r ia tes
a t the individ ual leve l ( i.e. indiv idua ls were ass igned va lues repre sent in g
s t ruc tura l contexts such as indus t ry codes , labor market segments , ne igh
borhoods) and inc luded as predic tors in s t ruc tura l models . More recent ly ,
m acro-leve l variab les in m ult i level m od els are m ea sur ed a s context-specific
aggr egate indices (e.g. m ean s or d ispers io n m easures ) of microleve l charac
terist ics or as global variables not expressed in term s of ind ivi du al ch arac
te r is t ics and inc luded in complex mul t i -equat ion sys tems or in dynamic
models permit t ing t ime-varying covar ia tes (DiPre te and Forr is ta l , 1994;
Mayer and Tuma, 1990) .
Educational a t ta inments have pers is tent , robust effects on la ter l i fe
outcomes across national contexts as well as in the U.S. But variat ions in
na t iona l pa t te rns have been observed because c ross - leve l mechanisms
vary . T w o catego ries of m ech an ism s have been observ ed: (1) the interac t ion
of individual t ra i ts with educational s tructures that vary across contexts
and (2) the l inkages am on g educa t iona l s t ruc tures and subse qu ent ach ieve
ment s tructures such as the workplace that vary across contexts . Educa
t iona l tracking sys tem s from e lem entary g rad es thro ug h t rans i t ions to jobs
from school have s tructural effects on individual pat terns of achievement,
as severa l essays in th is volu m e dem on s tra te (espec ia l ly those by G am ora n,
Ale xand er an d E ntwis le , W il lms , and Hal l inan) . The educat ion a l career is
a sequen ce of t ran si t ion s stratif ied over t ime by a relat ively orderly pa tter n
of cumula t ive advantage or d isadvantage wi th in school contexts . Across
school contexts heterogeneity and inequali ty are amplif ied by social c lass
segre gatio n further en han ced by differential o pp ortu nit ie s for school choice.
Fina l ly , school-to-job t rans i t ions culm inate wh at ap pea rs to be a na rro w ing
op por tun i ty s t ruc tu re fo r ach ievement wh ere d ive rgence ra the r than con
vergence is the dominant pa t te rn .
Parad oxica lly, there is evid ence that mo re loosely cou pled school-to-job
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 27/375
1
l inkages may produce grea ter he terogenei ty and inequal i ty than t ight ly
co up led system s. Kerckhoff s (1993) s tud y of school tracking an d the
transi t ion into the labor force in Great Bri tain provides s trong mult i level
evidence of the individual-school level interact ion and the school-to-job
linka ge. An d w hile Grea t Britain an d the
U.S.
differ systematically in some
aspects of t racking s tudents into academic versus vocational t racks, the
interface betw een scho ol and job is relat ively mo re am bi gu ou s in these tw o
coun tr ies than tha t observed in som e other contexts , par t icular ly Japan and
G erm any , leading to m ore var iance in w ork a nd ea rning s tra jec tor ies in the
former co untr ies . A m on g the la t te r a t ighter coupl ing be tw een ins t i tu t ional
contexts i s evident , i . e . a more de terminate t rack be tween curr icular /
c redent ia l v a l ida t ion in the educa t ional context and the ind ivid ual s ass ign
ment in the workplace exis ts . Direc t organiza t ional p lacements a re more
formal ly es tabl ished. A nd whi le the Japanese and G erm an sy s tems a re
di fferent in m an y othe r respects , the t ighter coup l ing of the edu cat io nal and
labor force contexts in these two co untries a pp ea rs to dim inish the level of
he terogenei ty and inequal i ty in the sa lary/wage sec tor fo l lowing educa
t ion . Ho w ever , DiPre te and M cM anus repo r t in th is volu m e tha t com par i
son of
U.S.
and G erm an earnings m obi l ity pa t te rns reveals h igher leve ls of
m ob ility in the U.S. In short, in the U.S., an d to a lesser extent in th e U. A.,
th e trac ks are no t as direct (Shav it an d Blossfeld, 1993). G ene ral cur ricu la—
even wi th abi l i ty t racking—appear to genera te more ambiguous pa ths to
jobs, especial ly in the absence of extended special ized educations, and
pe rh ap s to m ore rew ard ing , a lbe it m ore vola t ile , career earn ings pa t te rns .
In the employment context , work careers consis t of job transi t ion se
qu enc es that are anc hore d in ini t ia l inequali t ies s tem m ing from social c lass ,
s t a tus g roup o r educa t ion g roup membersh ip , bu t a re reor ien ted o r de
flected ov er t im e by the interact ion of ind ivid ua l w ork er and org aniz ation al
charac ter is t ics . Segmented labor markets provide d i f ferent ia l oppor tuni
t ies for achievement. Internal labor markets are conceptualized as frame
w ork s of cum ula t ive adv anta ge t radi tional ly associa ted wi th h igher e du ca
t ional or exper ient ia l c redent ia ls , la rger emp loym ent es ta bl ishm ents , pro
mot ional ladders , and wage and benef i t s t ruc tures encouraging long- term
at tachment pa t te rns , whi le secondary and cont ingent labor markets a re
divergent f rameworks of cumula t ive d isadvantage (see Kal leberg) .
Yet, w e are learning tha t these labe ls may obscu re cons iderable he tero
genei ty wi th in these ins ti tu t ional contexts . Spi le rma n a nd I shid a s s tu dy of
wo rk careers in la rge insurance com panies in the U .
S
and Japan repo r ts tha t
promotional schedules defy easy or overs implif ied views of fas t- tracks or
privi leged ascents . For example, they quest ion the general izabil i ty of
Ro senb aum s (1984) tou rna m ent m ode l of careers in h ierarchica l ly orga
nized em plo ym ent contexts across cul tura l se t t ings .
In addition, Kalleberg (in this volume) observes sufficient volatility in
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 28/375
labor markets during the recent period of economic reorganization and
globalization to inspire competing hypotheses regarding the polarization
of labor market structures. The view that current trends are towards a
bifurcated labor mark et structure in which som e trajectories can be labeled
as careers (i.e. as coherent job sequences w ith the accretion of hum an and
social resources), while others cannot as a result of the temporal and
substantive discontinuities between work episodes over the life course
hindering the accumulation of human capital. Changing employment
contexts in advanced industrial societies, and especially the U.S., are
threatening to attenuate the education-career link even further as trends
tow ards flexible specialization take hold and gain more ground. The
externalization of
labor
the disappearance of the lifetime
job
with a single
em ployer for growing segm ents of the workforce, and the reconstruction of
com pensations for work as we have know n them (chiefly employee ben
efits) are adding more volatility and further decoupling in the education-
workplace linkage (see Kalleberg in this volum e).
Continued divergence appears to result from the interaction of work
place structures and w ork ers ' characteristics for midlife status achievement
into retirement. Labor market duality, workplace mobility systems, rule
structures and institutionalized timetables (S0rensen, 1986; Spilerman,
1986;
Lawrence, 1986 are stratification systems tha t allocate rew ard s to jobs
(workers) of differential value to the organ ization. Jobs (workers) of higher
value are rew arded with wages and fringe benefits that provide incentives
for workers to remain attached to the organization and to sustain positive
trajectories of achievement. Jobs (workers) of lower value are rewarded less
generously and thus encourage more mobility, heterogeneity and flatter
trajectories of achievement. Multilevel approaches to these questions are
beg inning to link microlevel measures of ind ividua ls' career characteristics
with employer-level indicators of organizational structure, employment
policies, and other workplace characteristics to move towards a fuller
cross-level approach (Kalleberg, in this volum e).
A final m easure of life course achievem ent is economic sta tus at retire
ment. Diverging life course pathways within cohorts lead to increasing
economic inequality (O'Rand,
1995).
Gini coefficients of age group-specific
economic (income) inequality in the U. S. show greater inequality among
the elderly (age 60 and over) than any other adult age group (Crystal and
Shea, 1990). Comparisons with other countries reveal a similar pa ttern of
age-specific inequality, but also find country-level differences paralleling
those found regard ing school-job linkages (Hout in this volum e), and class
and earnings mobil i ty (DiPrete and McManus in this volume).
Political-economic structures, especially corporatist and social dem ocratic
welfare systems, are mediating environments that influence levels of in
equality and m obility by regulating institutional linkages and by differen-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 29/375
2
t ial ly redis tr ibuting resources to individuals following diverse careers
(Pam pel, 1993; Bu rkhau ser , Du nca n an d H au ser , 1994).
Relational pproa ches to Social Structure
W hile the t radi t ional ins t itu t ional appro ach to socia l s t ruc ture em ph a
sizes the inf luence of norm atively org aniz ed d om ain s of activity on career
trajector ies , ano ther appro ach ope rationalize s s truc ture as ne tw or ks or
sy ste m s of action (see Breiger, 1995, for a recen t review ). H er e the ach ieve
m en t process em an ate s from the access to, m atch ing , an d exch ang e of
resources among interconnected personal ne tworks . Accordingly, educa
t ional and labor market contexts of achievement are viewed as more than
socially-bounded and rule-dr iven domains; they are active systems of
interpersonal re la t ions tha t in terac t wi th normat ive sys tems and in many
w ay s t ranscend them . They incorpora te ac tors ' perspect ives mo re expl ic it ly
than tradit ional inst i tutional approaches, yet avoid an individualis t ic bias
by em be dd in g these persp ectiv es in relat ional system s w ith differential
global character is t ics . Thus, in the secondary school context, above and
be yo nd curr icular an d abil i ty grou ps, social t ies operate as social capital for
achievement . And in labor markets , workers ' ne tworks of in t ra- and
inter-organizational t ies have independent effects on careers that some
times supersede those of formal insti tutional contexts .
The s tu dy in this volu m e by Bid well and his associates on social ne tw ork s
and adolescent career development among 10th and 12th graders follows
this argu m ent . Ado lescent ne tw ork s are sys tems of resources tha t wor k to
orient s tudents ' behaviors towards different achievement trajector ies . Be
s ides ins t rum enta l ly useful resources , these netw orks prov ide socia l psy
cholog ical system s of su pp or t s ince they serve as au die nc es for de ve lop
ing self concepts.
The relational structure of the labor market also differentiates careers
above and beyond the cons tra ints of organiza t ional s t ruc tures and indi
vidual resources . While the t radi t ional ins t i tu t ional approach tends to
em ph as iz e the mo re formal ized or ra t ional ized l inkages bet w een no rm a
tive do m ain s, the relat ional appro ach p refers to exam ine f ields of s tr uc tur ed
re la t ionships among individuals tha t potent ia l ly cut across bounded do
m ain s. As such, it c laims to character ize social action system s mo re directly
and to preserve the perspectives of actors themselves. A recent argument
presented by Coleman (1991) following this approach is cr i t ical of career
analyses that character ize trajector ies of individuals within systems rather
tha n character izing sy stem s themselv es. H e conceives labor ma rke ts less as
struc tures tha n as processes, specifically m atch ing proc esses . These
proce sses consis t of system s of action in w hich w ork ers ha ve reso urces an d
preferences for selected job resources at the same t im e that em plo ye rs hav e
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 30/375
3
preferences for selected worker resources (including educational creden
tials) and hav e
job
resources
to
offer wo rkers. M atching
is
the produ ct of the
resources and preferences of the individual worker rel tive to 1) the re
sources of other w orkers in the market, 2 their interests in the
job,
and 3 the
distribution of job resources. The matching algorithm is a minimizing
function of the difference between
the
value of
a
work er's resources and the
value of the job resources held by the worker (Coleman, 1991: 6).
In this vein, Baron and Pfeffer (1994), whose work on organiza tions has
been central to the new structuralism in stratification research, have re
cently elaborated on the social psychology of organizations and inequality
following a relational appro ach to struc ture. They argue that work org ani
zations are com munities of fate, wh ere processes of social com pariso n,
ma tching, and evaluation are pervasive and consequential for ind ividuals
and for social structure alike. Work organizations are shaped by interac
tional and cultural processes permitting the micro-to-macro flow of influence.
Finally, a similar approach is taken by Burt (1992) who argues that
relational appro aches to structure pro vide the m eans for explaining relative
individual autonomy (agency) as well as institutional constraint. His con
cept of structura l ho les is central to the hypothesis that in com petitive
arenas like labor markets the optimal positions are derived from having
diverse ties to clusters which are themselves weakly tied. These structure s
of nonredundant ties increase the span of access to information and re
sources and optimize the wo rker's m arket pow er. Variations in access to
weak ties differentiate career trajectories, benefiting those with contac ts that
are more open an d inclusive. Bu rt's study of senior ma nagers in the volatile
environment of the high technology industry demonstrates the unique
contribution of the relational approach in transcending the boundaries of
traditional institutions to explain ind ividua l outcomes— as and w ell as the
possibility of systemic effects following from actions of individuals.
Perha ps the complem entarity of institutional and relational approa ches
for
the
stud y of career trajectories is most app arent in this vo lum e in Barbara
Heyns' interesting account of the social origins of entrepreneurship in
Post-Com mu nist Po land. Here the volatility of the transition from com mu
nism to a m arke t econom y reveals that social action with career outcom es
occurs in deinstitutiona lizing settings w ithin fields of social relations, but
that institutional origins have transcendent influences in the process.
Conclusions
The
new structuralism has benefited from impro ved m ethodologies an d
rapidly grow ing com parative databases that permit the direct examination
of cross-level effects. The new methodologies provide algorithms that m atch
individuals with social structures and, in some instances, incorporate time
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 31/375
explici t ly to capture social processes. They can more directly assess indi
vidual and s tructural outcomes as functions of micro and macro level
characteris t ics . The databases hav e imp rove d over the pas t four deca des in
a t leas t two ways : mic rodata sources hav e im prov ed w ith long i tudin al
informat ion on educ at ional and w ork career sequences tha t a re m or e a nd
mo re r epresen ta t ive . A nd , macroda ta bases have exp anded to ma tch ind i
vidual information to direct indicators of the contexts of their careers ,
inc luding employment organiza t ions , socia l ne tworks , and cross-nat ional
env i ronments .
Ins t i tu t ional and re la t ional theor ies provide complementary f rame
w or ks for the exam ination of cross-level effects . They pro vi de the con cep
tual tools to examine both insti tutional constraint and the relat ional bases
of ind ivi du al au ton om y. Acc ordingly, they enable us to acco unt for bo th the
heterogeneity of individual careers and their patterned differences. Their
new promise is to de l ineate the macro- to-micro and the micro- to-macro
flow of effects in social processes.
References
Alwin, Duane F. 1993. Aging, Personality and Social Change: The Stability of
Ind ividual Differences Over the Adult Life-Span. In
Life Span Development and
Behavior
edited by D. L. Featherman, R. M. Lerner and M. Perlmutter. Hillsdale
NJ: Erlbaum .
Baron, James N. and Pfeffer, Jeffrey.
1994.
The Social Psychology of Organiza tions
and Inequality. Social Psychology Quarterly 57:190-209.
Blau, Peter S. 1960. Structura l Effects.
American Sociological Review
25:178-193.
Breiger, Ronald L. 1995. Socioeconomic Achievem ent and Social Structu re. An-
nual Review of Sociology
21: forthcoming.
Burkhauser, R ichard V., D uncan, Greg J., Hause r, Richard.
1994.
Sharing Prosper
ity Across the Age Distribution: A Com parison of the United
tates
and Germany
in the 1980's.
Gerontologist
34:150-160.
Burt, Ronald
S.
1992.
Structural
Holes:
The Social Structure
o f Com petition. Cam bridge
MA: Harvard University.
Co lem an, James S. 1970. Properties of Collectivities. In
M acrosociology: Research
and Theory
edited by
J.
S. Coleman, A. E tzioni, and J. Porter. Boston: Allen and
Bacon:
1 102.
Coleman, James S. 1991. Matching Processes in the Labor Market.
Acta Sociologies
43: 3-12.
Crystal, Stephen and Shea, Dennis. 1990. Cum ulative Advan tage, Cum ulative
Disadvantage, and Inequality Among Elderly P eople.
Gerontologist
30:437-443.
Dannefer, Dale. 1987. Ageing As Intracohort Differentiation: Accen tuation, the
Matthew Effect and the Life Course.
Sociological Forum 2:211-236.
Dannefer, Dale. 1988. Differential Gerontology and the Stratified Life Course:
Conceptual and Methodological Issues.
Annual Review of Gerontology and Geri-
atrics 8: 3-36.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 32/375
15
Dannefer, Dale. 1991. The Race Is To the Swift: Images of Collective Aging. In
Metaphors of Aging in Science and the H umanities, ed ited by G. M. Kenyon, J. E.
Birren and J. J. F. Schroots. New York: Springer: 155-172.
DiPrete, Thom as A. and Forristal, Jerry
D.
1994. Multilevel Models: M ethods and
Substance. Annual Review of Sociology 20:331-357.
Easterlin, Richard. 1980.
Birth and
Fortune, New York: Basic Books.
Elder, Glen H., Jr. and O Rand, Angela M. 1995. Adult Lives in A Changing
Society. Pp . 452-475 in Sociological Perspectives on Social
Psychology,
edited b y K.
Cook, G. Fine and J. S. House. Boston: Allyn and Bacon.
Giddens, Anthony. 1976. New Rules of
Sociological
Method: A
Positive Critique
of
Interpretative
Sociologies.
N ew York: Basic.
Hen retta, John C.
1992.
Uniformity and Diversity: Life Course Institutionalization
and Late-Life Work Exit. Sociological
Quarterly,
33:
265-279.
Hogan, Dennis
O.
and N anM . Astone.
1985.
The Transition to Adulthood.
Annual
Review of Sociology 12:109-130.
Huber, Joan. (Ed). 1991. Macro-Micro Linkages in
Sociology.
New bury Park CA: Sage.
Kanter, Rosabeth M.
1977.
Men and Women
o f
the
Corporation.
New York: Basic.
Kerckhoff,
Alan C. 1976. The Status Attainment Process: Socialization or A lloca
tion. Social Forces, 55: 368381.
Kerckhoff, Alan C.
1993. Diverging
Pathways:
Social Structure and Career
D eflections.
Cam bridge University Press.
Kerckhoff, Alan C. 1995. Institutional Arrangem ents and Stratification Processes
in Industrial Societies. Annual
Review
o f Sociology,
21:
323-47
Kohli, M artin . 1986. Social Organization and Subjective Construction of the Life
Course. In Human Development and the Life
Course:
Multidisciplinary
P erspectives,
edited by A. B. S0rensen, F. E. Weinert and
L.
R. Sherrod. Hillsdale NJ: Erlbaum:
271-292.
Kohli, M artin. 1988. Ageing As A Challenge for Sociological Theory. Ageing and
Society 8: 367-394.
Lawrence, Barbara S. 1984. Age-Grading: The Implicit Organization Timetable.
Journal
of
Occupational Behavior 5: 23-35.
M addox, George L. 1987. Aging Differently.
Gerontologist
27:
557-564.
Mayer, Karl U. and Schoepflin, Urs. 1989. The State and the Life Course . Annual
Review
of
Sociology 15:187-209.
Mayer, Karl U. and Tum a, Nancy B.
(Eds). 1990.
Event History In Life Course
Research.
University of W isconsin.
Merton , Robert R. 1968. The Matthew Effect in Science. Science 159 (January 5):
56-63.
Mills,
C.Wright. 1953.
Two Styles of Research
in
C urrentSocial
Studies. Philosophy
of Science,
20: 266-275.
Modell, John. 1989.
Into
One s
Own:
From
Youth
To Adulthood
in the
United
States
1920-1975. University of California.
O Rand, Angela M. 1995. The Cumu lative Stratification of the Life Course. In
Handbook ofAging
and the Social
Sciences.
Fourth
Edition,
edited by R. Binstock an d
L. R. George. NY: Academic
Press:
forthcoming.
Pam pel, Fred C. 1993. Relative Cohort Size and the Easterlin Effect.
American
Sociological
Review,
58:496-514.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 33/375
16
Riley, Matilda White. 1993. The Com ing Revolution in Age Structu re. W orking
Paper R-110, Pepper Institute on Aging and Public Policy, Florida State Univer
sity, Tallahassee FL.
Rindfuss, Ronald R. 1991. The Young Adult Years: Diversity, Structural Change
and Fertility.
Demography
28:493-512.
Rosenbaum , Joel. 1978. The Structure of Opportunity in School. Social Forces 57:
236-256.
Rosenbaum, Joel. 1984.
Career Mobility in a Corporate
Hierarchy. NY: Academic.
Rosenfeld, Rachel. 1992. Job Mobility and Career Processes.
Annual Review of
Sociology
18:39-61.
Sorensen, Aage B. 1986. Social Structure and Mechanisms of Life Course Pro
cesses. In
Human Development and the Life
Course:
Multidisciplinary
Perspectives
edited by A.
B.
Serensen,
F.
E. W einert, and L.
R
Sherrod. Hillsdale NJ: Erlbaum:
177197.
Spilerman, Seymour. 1977. Careers, Labor Market Structure, and Socioeconomic
Achievement.
merican journal o f Sociology
83: 551-593.
Spilerman, Seymour. 1986. Organ izational Rules and the Features of Work Ca
reers. Research on Social Stratification and Mobility
5: 41-102.
Wilensky, Harold L. 1981. Family Life Cycle, Work and the Quality of Life:
Reflections on the Roots of Happiness, Despair and Indifference in Modern
Society. In
Working Life
edited y B. Gardell and G Johansson. New
York;
Wiley:
235-265.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 34/375
Social Psychological
Aspects of Achievement
eylan
T
Mortimer
University
of innesota
Understanding the social psychological predictors of attainment was a
major preoccupation of status attainment research in the late sixties and
seventies. Sociologists examined the social and behavioral anteceden ts of
achievement aspirations, such as family socio-economic background, sig
nificant others influence and academ ic performance, as we ll as the conse
quences of these a ttitude s for a ttainm ents (Sewell, et
al, 1969;
Sewell, et a l ,
1970; Hauser,
1971;
Gordon, 1972). Clear l nks were demonstrated between
educationa l and occupational origins (as indicated by fathers achieve
ments) and adolescents aspirations. Parental encouragement and school
performance largely m ediated these effects Kerckhoff, 1974; Hauser, 1971;
Duncan, et a l, 1972: Alexander, et al.,
1975;
Wilson and Portes, 1975).
This body of work, focused on adolescence and early adulthood in the
United States, laid the groundwork for subsequent elaborations of the
status attainm ent m odel which have much enriched our conceptualization
and un derstand ing of the social psychology of achievem ent. Social scien
tists have s tudied a wide array of achievement-related attitudes and moti
vations, diverse attainment outcomes, socialization processes, and sub
group differences. Cross-national studies have alerted researchers to struc
tural divergences in processes of achievement-relevant socialization and
allocation. Finally, investigators have extended this research trad ition to
exam ine the social psychological dynam ics of achievem ent thro ughout the
life course.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 35/375
S
Expansion of the Psychological Precursors
and ttainment Outco me s
The ear ly at tainment researchers mainly focused on aspirations and
m or e realis t ic pl an s w ith respect to future edu catio n and occ upa tion. But
as Spermer and F eath erm an's (1978) review of the l i terature on ach ieve
m en t am bit ion s show ed, a m uch wide r ar ray of psychologica l d imen s ions
are imp licated in achiev em ent. In this reg ard , sociologists ha ve mu ch to
learn from cognate social science disciplines.
Econom is ts speak of tas tes for em plo ym en t an d the pro pen s i ty to
w or k or , converse ly the prop ens i ty for un em plo ym ent , the tas te for
lei su re (Co rcoran an d Hill, 1980), an d the tast e for va ca tio n (Coe , 1978)
as source s of occu pation al outc om es. They dr aw attentio n to at t i tud es
rega rdin g the reservat ion w age , tha t i s the m ini m um offered w ag e
necessary to induce an un em plo ye d person to take up em plo ym en t (Hui ,
1991:1341) , f inding, not surpr is ingly , that longer spells of em pl oy m en t are
associated with higher minimally acceptable wage offers (Hui, 1991; Baker
and Elias, 1991).
Psychologists point to work motivation, the need for achievement,
expecta t ions about the consequences of achievement- re la ted behaviors ,
an d perce ptio ns of the op po rtun ity s tructu re. For exa m ple, in a factor
analytic s tudy, Feather (1986) reported two dimensions of or ientation
tow ard future job prospec ts am on g Austra l ian secondary school s tud en ts .
The first, a value factor, included indicators of interest in work, job need,
an d job w an t. The second w as an expectation factor which he called
un em plo ym en t disa pp ointm ent . This cons truct ref lec ts the deg ree of
helplessness or pessimism about job prospects , including the level of
confidence ab out finding
job, the difficulty of doi ng so, the am ou nt of t ime
it will
take,
an d the sense of personal control regarding em ploy me nt outcom es,
Self-conceptions as sources of achievement are of interest to psycholo
gists as well as sociologically-oriented social psychologists (see Spermer
an d Fe athe rm an, 1978; Schw albe and G ecas, 1988). Inve stigato rs ha ve
noted pos i t ive re la t ionships be tween se l f -es teem and educat ional p lans
(Kerckhoff, 1974), as w ell as be tw ee n self-esteem a nd o cc up atio na l asp ira
t ions (Gordon , 1972). Spermer an d Otto (1985), in their pan el s tud y of you ng
people in Washington s ta te , found tha t women with higher se l f -es teem,
m ea su red wh ile s ti ll in high school, exp erienced fewer mo nt hs of sub se
quent unemployment over the fol lowing thi r teen-year per iod; however ,
there was no such effect for men.
Whereas self-esteem refers to the global sense of worth, another set of
self-conceptions m ay be even m ore per t in ent to achie vem ent. Internal
control orientation, mastery, self-efficacy and the sense of competence all
reflect expectations about the likelihood of successful goal attainment.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 36/375
9
Bandura (1988) summar izes a la rge body of research demonstra t ing the
importance of self-efficacy for work-related behaviors. A study of a panel
of 1,000 nin th -g rad er s in
St.
Paul , M innesota provides su pp or t for B and ura ' s
form ulatio n (see Call , e t al ., 1993). A n attem pt w as ma de to dete rm in e th e
socia l psychologica l anteceden ts of edu cat ional success amon g you th w ho
come from poor families ( those whose family incomes fall below the
po ve rty l ine) . Fo ur years after the s tu dy beg an, those po or yo uth w h o
ap pe are d to be on a successful edu catio nal trajectory w ere sepa rate d
f rom those w h o w ere not . The successful yo uth ha d not dr op pe d ou t of
school, they ha d a
B
ave rage or abov e, an d they had ta ken som e action to
pr ep are for college adm ission (e.g., taken the Scholastic A chie vem ent Test ,
talked to a counsellor about college, sent for , or submitted, a college
ap plic atio n, or oth er concrete activities). T here w er e no significant differ
ences between the two groups in aspira t ions or p lans for educat ional
attainment, nor were there differences in their occupational aspirations.
However , the more successful poor adolescents had a s ignif icantly higher
sen se of eco no mic self-efficacy. Fu rth erm ore , the re is ev ide nc e tha t self
rate d confidence in f inding a job is predic tive of em pl oy m en t success after
leaving secondary school (Feather and O'Brien,
1986;
O'Brien an d Feather ,
1990).
S om e s oc i a l p s yc ho log i s t s f oc us on oc c upa t i ona l r e wa r d va lue s
(Rosenb erg, 1957; D avis , 1964) . A pane l s tud y of U niversi ty of M ichigan
me n (Mo r t imer , e t a l , 1986) sh ow ed that em ph asis on extr insic occu pation al
values whi le s t i l l in col lege predic ted employment s tabi l i ty and income
at ta inm ent a dec ad e following gradu at ion. In add i t ion to high er incom es ,
those seniors wi th higher extr ins ic va lues—who placed grea ter emphas is
on mo ney , advancem ent , p res t ige and securi ty— had le ss u nem ploy me nt ,
invo lun ta ry pa r t - t ime emp loym ent and un derem ploy m ent , a s we l l as
fewer changes in career direction during the ten years following college
gra du atio n. Earl ier interest in the intr insic rew ard s of w or k pr edi cted
au ton om y on the job a decad e la ter; and p eople-or iented values— such as an
interest in working with people and being of use to the society—fostered
ad ult wo rk that w as hig h in social conten t. M ore recent research on high
school s tud ents sh ow s tha t the i r occupat ional va lues increas ingly pred ic t
the features of their par t- t ime jobs as they move through high school
(Mortimer , et al . , 1994) . Ninth-grade occupational value or ientations had
no signif icant effects on tenth-grade work experiences. However , intr insic
values me asu red in the tenth grade ha d a pos i t ive effec t on the op po r tun i ty
to learn job skills in the eleven th; the pa tter n w as replicated sub seq ue ntly ,
betw een the junior and senior years . M oreover , in t r ins ic va lues m eas ure d
in the sophomore and junior years s ignif icant ly predic ted subsequent
op po r tun i t ies to be helpful to others a t wo rk.
Of great interest f rom a social policy perspective are the at t i tudinal
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 37/375
2
precursors of unemployment and career instabil i ty (Mortimer , 1994) . For
exa m ple , a m eas ure of adh erenc e to the Protes tant w ork e thic w as fou nd
to dis t in gu ish s tu de nts w h o found jobs one year (Feather an d O'Brien, 1986)
an d tw o years (O'Brien and Feather , 1990) after leavin g A ustra lian seco nd
ary schools . Grea ter perceived need for a job also chara cter ized those w h o
w ere m ore successful in f inding w or k (Feather and O 'Brien, 1986) .
There is evidence tha t young people who are more dis t ressed have
grea ter diff iculty in the job m ark et af ter leaving school. U ne m pl oy ed yo un g
pe op le in Au stralia ha d low er life satisfaction and m anifested grea ter s tress ,
a less posit ive at t i tu de, an d m ore dep ressiv e affect t w o years ear l ier w h en
they were s t i l l in school than those who found employment af ter leaving
scho ol (Feathe r an d O'Brien, 1986; O'Brien a nd Fe athe r, 1990). In an ot he r
A ustra l ian s tud y (Winef ie ld an d Tiggem ann, 1985), yo un g peop le w ho
became unemployed a l so expressed more boredom and depress ion , and
ma les w ho became un em plo ye d expressed mo re lonel iness, before leaving
school.
At ta inm ent researchers in i tial ly focused on edu cat ional a t ta in m ent and
occupational prestige ear ly in the career as indicators of socio-economic
stat us ou tcom es, and to a lesser extent, incom e. A chiev em ent-relate d
at t i tude s have recent ly been associa ted w ith par t icular labor mark et entry
segments predictive of divergent occupational career l ines (Ashton and
Sung, 1991) . Other s tudies have l inked pr ior psychological at tr ibutes with
a diverse ar ray of o ther occupat ional ly- re levant co nsequen ces w hose s ig
nif icance extends throughout the work career—such as work autonomy
(M ortim er, et al. , 1986), job stresso rs (S han aha n, et al., 1991), su bs tan tiv e
com plex ity of tasks an d other s truc tural imp erativ es of the job (Kohn an d
Schooler, 1983).
These examples i l lus t rate the m ul t i tu de of achievem ent- re la ted a t t i tude s
and a t ta inment- re la ted outcomes tha t have been scrut inized by inves t iga
tors f rom a w ide array of social science disciplines. U nd ers ta nd in g the
sources of these diverse achievement- re la ted mot ivat ions and a t t i tudes
constitutes a major challenge for future research.
The Socialization of chievem ent Orientations
The re la t ionships be tw een socia l c lass and chi ld educa t ional an d occup a
tional ou tcom es ha ve been found to be m ed iate d by social psych ologica l
dimensions, such as the favorableness of the self-concept, the sense of
control of the environment , and aspira t ions , as wel l as by educat ional
performance (Kerckhoff,
1972).
H ow do social class , gen der , family c om po
sit ion, parent-child relat ionships, peer relat ions, and school experiences
lead to the development of those at t i tudes and values that predict achieve
m ent? Gecas (1979) ha s do cu m en ted l ink ages be tw een social class and
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 38/375
2
m ode s of parental control, pow er relationships in the family, and com mu
nication betw een paren t and child. These differences in parent-ch ild
relations fostered achievement-related psychological advantages or disad
van tages in children. Kohn (1969,1981) speculated that
class
differences in
parental values with respect to self-direction and conformity limit
intergenerational m obility. That is, emp hasis on conformity am ong blue
collar workers would promote values and behaviors in children that are
ada ptive for similar work, bu t wou ld not equip them w ith the self-directed
orientations and behaviors facilitative of success in managerial and profes
sional
roles.
Feather (1986) found that high school stud en ts of high er social
class origin believed that unem ployment w as attributable to more internal
causes, and this belief was negatively related to helplessness and pessi
mism.
However, early studies examined the socio-economic precursors of
educational and occupational aspirations, under the assumption that the
status of the family of origin was fixed. Given grow ing attention to
mfragenerational mobility, the faultiness of this assumption has become
increasingly a ppa rent. That is, paren tal socio-economic statu s may chang e
dur ing the periods of childhood and adolescence, a time when imp ortant
vocational socialization is taking place (Featherman and Spenner, 1988).
Family socio-economic mobility may foster red irection in child ren s trajec
tories of achievem ent.
We know little abou t the sources of occupational value s, though the re is
some evidence that they are transmitted intergenerationally throu gh close
and communicative parent child relations that promote identification and
modelling. The closeness of the father-son bond fosters intrinsic values in
professional families and extrinsic values in business families (Mortimer
and Kumka,
1982).
Supportive relations with fathers also stimu late higher
levels of anticipated wo rk involvem ent (Mortimer, et al., 1986). On going
research suggests that the link between socio-economic background and
occupational rew ard v alues emerges during the high school years. W ork
value s become m ore s trongly linked to socio-economic origins, in
a
manner
characteristic of adu lts, as students m ature. That is, socio-economic back
grou nd has no significant impact on boys and girls occupational values in
the 9th
grade.
Howe ver, by the senior yea r it had significant positive effects
on intrinsic value s (M ortimer, et a l , 1994).
It is widely assumed that the influence of high school track placement
(e.g., academic, vocational, comm ercial,
etc.
and ability group ing on achieve
me nt orientations is m ediated by socialization processes, especially teacher-
stude nt and peer relationships. An opposing, thoug h not necessarily
contradictory, point of view is that students leam of their likely occupa
tional destinations thro ugh observation of these self-placements, as well as
the fates of their similarly-situated predecessors; they form aspirations
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 39/375
accordingly (see Gamo ran, this volum e). Though such socialization and
allocative mechanisms probably work in tandem, understanding their
relative importance, or the conditions that make one or the other process
more salient, wo uld elucidate im por tant social psychological dynam ics of
achievement.
Adolescents ow n wo rk experience may also be implicated in the social
psychological process of achievement. An ongoing study of high school
students in St. Paul, Minnesota, shows that positive work experiences—
especially the opportunity to learn skills at work which are perceived as
useful in the future—induce intrinsic occupational values (M ortimer, et al.,
1994). At
the
same time, positive connections between school and w ork and
advancement opportunity enhanced boys sense of mastery (Finch, et al.,
1991); girls self-perceived efficacy increased when they reported that they
were paid well. Feather
1986)
similarly finds that eleventh grade students
wh o had wo rk experience prior to leaving high school had lower levels of
helplessness and pessimism in relation to future job prospects. The positive
association between adolescent part-time w ork and both em ploym ent and
income in the years immediately following high school (Millham, et al.,
1978; Freeman and W ise,
1979;
Meyer and W ise,
1982:
Mortimer a nd Finch,
1986) may be at least partially attributable to these salutary psychological
outcom es. Study of the St. Paul you th s transition to adultho od may
prov ide direct evidence with respect
to
such processes of social psychologi
cal mediation.
Finally, as Bidwell, Plank and Muller (this volume) point out, whereas
investigators have extensively examined the influence of significant oth
ers— including pa ren ts, teachers, and peers—on asp irations, little is known
about the ways in which the organization of adolescents networks of
associations influence aspirations
as
well
as
other orientations
to
achievement.
Subgroup Differences
in the Social Psychology of chievement
In another kind of elaboration of the status a ttainment mo del, research
ers are increasingly attending to subgroup differences in the process of
attainment. Whereas early studies were often based on men, researchers
increasingly inquired about gender-specific attitudinal p redictors of attain
men t. Though em ploymen t is increasingly comm on among adu lt wom en,
adolescent girls still encounter traditional values emphasizing the impor
tance of appearance, popularity, marriage and parenthood which may
interfere with achievement-related effort and occupational advancement.
Bidwell et al. (this volume) present ev idence that cohesive peer netw orks
strengthen bo ys ambitions, while reducing those of girls. Moreover, by
observing the experiences of their mothers and other adult women, girls
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 40/375
3
ma y become inc reas ing ly awa re of the dilem m as an d role conflicts inv olv ed
in com bin ing fam ily l ife and wo r k , as wel l as the s ex-typed job m arke t and
other obstacles
to
wom en s ach ievement . High occupa t iona l a t t a inmen ts
m ay co me to be seen as incom patible w ith the achiev em ent of familial goals .
Th ou gh th ere has been m uch controve rsy abou t causal order , it is fairly w ell
established that ,
for
w o m e n ,
a
hig h level
of
f amily invo lvem ent
and
large r
family size
are
associated with lower levels
of
female socio-economic
achievement (McLaughl in , et al. 1988).
Gender differences in educa t iona l and occupat ional p lans and a spira
t ions cont inue to be of interest. W hile ear ly s tudie s— in the s ixties and
before— show ed tha t boys have higher achievem ent or ienta t ions than gir ls ;
s tud ies
in the
sevent ies
and
e ighties indica ted
no
consis tent ge nd er differ
ence .
M ore recent research sugges ts that ado lescen t gir ls of ten sur pa ss bo ys
in
the
level
of
the i r aspira t ions . S tudy
of a
r epresenta t ive panel
of
ado les
cents drawn f rom
the St.
Paul pub l ic schools
is
consis tent with this lat ter
t r end (Dennehy and Mortimer, 1993). It s hows tha t, in the agg rega te, gir ls
have educat ional aspira t ions and plan s, as well as occu pation al asp iratio ns,
that exceed those
of
boys . M oreover , the i r com mitm ent
to
w o r k
in the
future—that is , the impo r tance w hich they a ttach to future w ork— is eq uiva
lent to that of boys . However , the i r sense of eco nom ic efficacy— their
pred ic t ions abo ut
the
l ikelihood
of
their actually bein g able
to
achieve their
econom ic goals , such as ow nin g the ir ow n ho me or hav ing a job wh ich pa ys
wel l—is , on the ave rage , s ignif icantly low er than th ose of boys (Dennehy
and Mo r t imer ,
1993).
Cons is tent ly , econom is ts find tha t w om en h av e low er
reservat ion wages than
men a
pa t tern tha t w ou ld de pres s the i r wag es
relative
to men
( Hu i , 1991; Lynch, 1983; Jon es, 1989).
Severa l s tudies demonstra te tha t g i r ls have a weaker sense of self-
efficacy than boys
in
general (Maccoby
and
Jacklin, 1974; Simmons
and
Blyth,
1987;
Gecas ,
1989:
p .
305;
F inch,
et
al.,
1991).
If in
fact,
it
is
not
so m uch
aspira t ions tha t s t im ula te pers is tence in achievem ent- re la ted behavior , but
a real sense of efficacy and confidence tha t on e s efforts are likely to be
successful, girls
may
still
be
psychologica lly disadv anta ged desp i te the i r
high aspiratio ns. Ad olesce nt girls ha ve also bee n found
to
h a ve l owe r
self-
es teem than boy s (S imm ons and B lyth , 1987) and high er levels of de pres s ive
affect (Rutte r,
1986),
attr ibutes w hich m ay also imp ed e success in the jo market.
Whereas these s tudies , l ike most in this f ield, generally report mean
differences, by g e nde r , in achievement- re levant a t t i tudes and p sycholog i
ca l t ra i ts , such ma y obscure im por ta nt d i fferences am on g w om en
in
or ien
ta t ions and pl an nin g s t ra tegies. To w ha t extent do you ng w om en view their
futures as con t ingen t on fu ture spou ses , chi ldren, and o thers , d im inishin g
their pro pen s i ty to m ake f irm plans given unp redic tabi l i ty in t he need s and
d e m a n d s
of
oth ers (Ha gestad , 1992)? Geissler
and
K rug er (1992),
in
their
qual i ta t ive s tudy of Germ an you ng wo me n, no te d if fe ren t m ode ls of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 41/375
4
biograp hica l cont inu i ty . Some w om en cons t ruc t the i r fu ture b iogra phies
w ith reference to career, focusing o n obtainin g professional qualif ications
an d delay ing marr iage . O the rs , in contras t , expect to enact t rad it iona l
female adult roles , emphasizing l imited labor force part icipation and
econom ic depe nd enc e on a hu sba nd . Still o thers cons ider job an d family to
be equal ly im porta nt . Such divergence in p lann ing mo de s wo uld cer ta in ly
influence the intensi ty of achievement-related s tr iving and the extent of
future a t ta inment .
Racial an d ethnic differences in the social psyc holog y of ach ieve m ent are
of con siderab le impo rtanc e. O gb u (1989) f inds that des pite hig h as pira
t ions, black you th are easi ly disco urag ed in their a t te m pts to achieve. Giv en
the long his tory of racial discrimination in the United States , many black
youth believe that they wil l confront obstacles in their quest for occupa
t ional and economic at tainments even if they are successful in the educa
tional system. Instead of pressing forward in the face of difficulties, their
folk cultur e of succ ess leads them to assu m e a so m ew ha t fatal is t ic
po stu re , W ha t 's the use of t ryin g? In this w ay , Black yo uth are chann elled
away from schoolwork and toward behaviors tha t d iminish the l ike l ihood
of legit ima te em plo ym en t success . Indicating the real ism of these at t i tu de s ,
Wilson (1987) and Sull ivan (1989) document the absence of employment
opportunit ies for young people in the inner ci ty.
M uc h m ore a t tent ion needs to be d i rec ted to sub gro up di f ferences in the
social psychology of achievement, and to the sources of these variat ions.
Giv en contin uin gly ra pid ch ang e in gen de r roles , there is ne ed for careful
monitoring of gender differences in achievement orientat ions and their
conseq uence s . W hat exper iences de term ine wheth er g i r ls wi l l be o r iented
to careers or to m ore trad it ional gend er roles? M oreo ver, li t tle is kn ow n
abo ut the achievem ent or ienta t ions of m inor i ty gro ups o th er than Blacks.
Are the achievement a t t i tudes which predic t a t ta inments in the genera l
po pu la t io n a lso character is t ic of h igh-achieving Asian-Am ericans? A nd
what about very recent newcomers , such as the Hmong, who exhibi t h igh
asp irat ion s as well as factors m it igatin g again st their achieve m ent— fam il
ial po ve rty as well as early m arriag e and fert il ity in the adolescen t gen era
tion (McNall, et al., 1994).
Structural Moderation of the
Social Psychological Dy nam ics of ttainment
There are differences in the process of at tainment depending on the
sector of the labor market , whether in the primary or core, or in the
periphery or secondary sector (see Beck, et a l . ' s discussion of income
attain m en t, 1978, in the U.S.; an d A shton and Su ng 's , 1991, assessm ent of
yo un g peop le 's early careers in Bri tain). H ow eve r, s tud ies of labor m ark et
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 42/375
5
segmenta t ion and a t ta inment have genera l ly ignored soc ia l psychologica l
aspects of the at ta inm ent p rocess . Kerckhoff (1989) sug gests tha t the s tatu s
a t ta inm ent and labor segmen ta t ion mo dels be in tegra ted to dev elop a mo re
balanced p ic ture .
Co m par a t ive , c ross -na t iona l s tudies of the soc ia l psycho logy of achieve
ment would much con t r ibu te to our unde rs tand ing o f s t ruc tu ra l ly -de te r
m ine d var iabi l ity in the process . W hereas in the Un i ted Sta tes educa t ion a l
at ta inment is measured by years of formal schooling and the receipt of
dip lomas and degrees , of much grea ter importance in Bri ta in i s achieve
m en t in the com plex system of qualif icat ions Kerckhoff, 1990). The se ar e
obta ined through diverse exper iences , inc luding regular school ing , ap
prenticeship, on-the-job training, vocational or technical t ra ining and uni
vers i ty- leve l edu ca t ion , an d are cert if ied by examinat ion . M any s tud ents ,
especial ly m en, wh o left school early w ith low-level qualif icat ions, and w h o
ha d been in the less favored secondary school loca t ions , obta ine d ad di
t iona l qua l i f ica t ions throug h these opp ortun i t ies . At the sam e t ime, fur
ther edu ca t io n is a re la tive ly s t rong predic to r of occup at iona l p res t ige
a t ta inm ent m eas ure d a t the age of 23 . Th us , m any Bri ti sh you ng peop le a re
able to compensate for their ini t ia l disadvantage in the highly s trat if ied
sp on sor ed mo bi l i ty sys tem of regular schooling by par t ic ipa t ion in the
a l te rna t ive ro ut e (p . 160).
Kerckhoff (1993) f inds s tructura l sourc es of cu m ula tive edu catio nal
adv ant age in the Brit ish educa t iona l sys tem. Som e s tud ents mo ve t hr ou gh
high -ach ievin g c areer l ines , i .e ., junio r schoo l hig h abil i ty g ro up s, e l ite
secon dary schools , and h igh er educ a t ion . O thers mov e from a junior school
low abi l i ty group, to a low abi l i ty group in a comprehens ive school , and
terminate the i r educa t ion wi th no pos tsecondary courses . He concludes ,
at each s tage in the educational career. . . there were very s trong effects of
organ iza t iona l s t ruc ture on the achievem ents of the indiv idua l s tuden ts , ne t
of the effects of the s tudents ' social origins , personal characteris t ics , and
past experien ces an d perfo rm ance s. (p. 200)
Kerckhoff 's com para t ive s tudies of the t rans i tion to ad ul th oo d in G rea t
Bri ta in and the United States , though focused on s tructural differences ,
br ing to mind severa l very in te res t ing ques t ions about psychologica l dy
nam ics . W hat indiv idua l - leve l consequences— cogni t ive , a ffec tive , an d
behav io ra l— are fostered by regu la r and by a l t e rna t ive educa t iona l
s t ruc tu res an d al loca t ive me chanism s? For exam ple , p lacem ent in low er-
leve l school ing in Bri ta in (e .g ., seconda ry m od ern as op pos ed to gra m m ar
schools) given the highly public nature of ass ignment, could detract from
self-esteem, the m otiv atio n to lea m an d the ex pe nd itur e of effort . School
typ e (e.g. , gr am m ar school, com prehe ns ive school , e tc .) and abil ity gr ou p
ing hav e s ignificant effects on the level of edu catio na l qualif ications eve n
tually ob taine d, even w hen abil ity, as gau ged by objective tes t as well as by
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 43/375
6
self an d teacher ratings, is controlled. W hether this results from variation
in teachers' expectations, self-labelling and adjustment of aspirations or
other processes, are questions remaining to be addressed. Moreover, one
might investigate the psychological consequences of the alternative route.
Might this opportunity reduce the alienation produced by assignment to
lower level secondary schools, or to lower ability gro up s, enhancing
commitment to both regular and further education amo ng those wh o
are not destined to receive higher educational degrees?
There has been little attention to differences in culture, social structure,
or the institutional linkages between school and work that could lead
certain psychological dimensions
to
be more important for achievement in
one context than another. Such studies would yield un der stan ding of the
interactions of structure s and attitudes in predicting attainm ent ou tcom es.
For example, the level of institutional connection between school and work
cou ld importan tly m oderate the effects of achievem ent orientations. The
United States is characterized by loose institutional connections; Germ any
and Japan represent countries w ith tighter, more highly structured linkages
(see H am ilton,
1990;
Rosenbaum , et a l , 1990). Because of clear institution
alized connections between particular schools and firms in Japan, the
student's prior academic achievement and teacher recommendations are
highly consequential for obtaining placement in high-quality jobs
(Rosenbaum, et al, 1990). In such circumstances, psychosocial variables
that influence school achievement could have greater relevance for early
attainment outcomes among non-college youth than in the United States,
where employers pay little attention to recent school leavers' academic
performance, and schools have few resources to respond to requests for
transcripts and other records, even when em ployers want them (Borman,
1991: 41).
These comparative analyses provide insight into possible structural
bases of the prob lems of you th in the United States. Th ough social
psycho logists generally look to the microcontexts of develo pm ent (e.g, the
family, the school classroom, or the peer group) as the central d eterm inan ts
of achievem ent-related orientations, behav iors and o utcom es, these cross-
national studies draw attention to what might be important structural
underpinnings of the much-heralded
U.S.
weaknesses in educational per
formance and economic productivity. he system of further educatio n in
Britain, like the institution of apprenticeship in Germany, and the struc
tured linkages between secondary schools and industrial corporations in
Japan, provide clear economic incentives for the acquisition of job-related
know ledge an d skills prior to an d /o r following the completion of second
ary education. Com parable structures prom oting non-college you th's
continuing motivation to learn and to enhance their human capital are
generally lacking in the United States. American non-college youth, in the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 44/375
7
years im mediately after high school, drift betw een jobs whose skill levels
are the
same,
or quite similar to , the jobs they were able to obtain when they
were still in high school (Osterman , 1989). N or are they likely to be enrolled
in the kinds of highly focused, specialized programs (like the system of
further e ducatio n opp ortun ities in Great Britain) after full-time labor
force entry that increase their economic prod uctivity . No t surp risingly,
they have become increasingly disadvantaged economically and socially
(W. T. Grant Fou nda tion, 1988).
The Social Psychology of chievement
through the Life Course
Most studies of the achievement process have focused on you ng pe ople.
Researchers typically survey aspirations, values, and o ther orientations of
students prior to leaving full-time education . Educational sociologists
focus on school-related determinants (such as tracking, ability grouping
and external rewards) of intrinsic motivation and effort in school, which
influence academic achievement and educational attainmen t (see chapters
by Hallinan an d Gamoran , this volume). Those interested in career deci
sion-making have assessed the connection between students' values and
occupational choices (Davis, 1964; Mortimer and Kumka, 1982; see also
He yns , this volum e, for assessmen t of laissez-faire values and proprieto r
ship aspirations am ong Polish business students.) ome investigators h ave
followed youthful panel members over a period of time to ascertain the
predictive capacity of these psychological attributes for educational and
occupational attainments early, and sometimes later, in the work career
(Mortimer, et al., 1986; Clausen, 1993).
In the early status attainme nt stud ies, aspirations were often m easu red
at one point in time, and assumed to be relatively invariant over time.
However, Kerckhoff's study (1974) suggested that adolescents form their
educational and occupational aspirations du ring the high school years, and
that aspirations were not as stable as assumed in the status attainment
literature. Our stud y of adolescents in St. Paul found that both bo ys ' and
girls' educational aspirations and plans actually declined as they moved
throug h high school (Dennehy and M ortimer
(1993).
The same pattern was
true of their occup ational aspirations. These decreasing levels of a spiration
may indicate growing realism as adolescents become more aware of labor
marke t realities. An other study indicates instability in aspira tions in the
years after high school as well. Rindfuss and h is colleagues (1990), usin g
National Longitudinal Survey data from 1972, the resp on dents ' last year of
high school, and follow -ups in 1973,1974,1976, and
1979,
found that less than
one qua rter of the youth h ad the same occupational expectation at all time
period s. We need to know m uch more about the sources of stability and
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 45/375
8
change dur ing th is format ive per iod . We a lso need more s tudy of the
processes of crystal l izat ion of achievement-related orientat ions as young
peop le m atu re . At w ha t poin t do person s becom e aw are tha t there a re
intrinsic and extrinsic dim ens ion s of w ork , each of wh ich can be re w ar di ng
and that these rewards are differential ly available in various occupational
roles? Do these increasingly soph is t icated con ception s of w ork ind uce
change in occupat iona l aspi ra t ions as young persons acquire more labor
force experience?
On ce a per son has entered the workforce , em plo ym ent an d joblessness,
as well as the quali ty of w or k (M ortimer e t a l ., 1986; Ko hn a nd Schoo ler,
1983), can influence div erse psycholog ical a t tr ibute s that influence sub se
qu ent occu pat iona l success. The ear ly w ork career m ay be espec ia lly
important in this regard, given that work orientat ions are in a part icularly
forma t ive s tage . If ear ly jobs a re uns tab le an d unre w ard ing , w ork ers m ay
become a l iena ted f rom work, develop poor work habi ts , and become less
at tach ed to the workforce (Corco ran an d Hil l , 1980:40). Experiences of early
joblessness could foster the development of psychological characteris t ics
that affect the likel ihood of future em plo ym en t. There is evid enc e th at
unemployment fos te rs d is t ress , se l f -b lame and o ther psychopathology
(Hamilton, e t a l . ,
1991;
Kessler, e t a l. , 1989) wh ich co uld jeo pa rdi ze sub se
qu ent success in the labor m arket . Un em plo ym ent a lso dep resses the
rese rva t ion w age (Hui , 1991). F requen t a n d /o r pe rs i st en t un em plo ym ent
m ay cause work ers to becam e discourage d, dr op pin g out of the labor force
bec ause they believe that wo rk is no t available (DiPrete , 1981). U ne m plo y
m en t, in fact , m ay al ter tas tes , ski l ls , an d m otiva tions in such a w ay tha t it
is pe rp etu ate d over t im e (Baker and Elias, 1991).
Compara t ive h i s to r i ca l s tud ie s o f l ives emphas ize the inc reas ing
destrucrurat ion and individualizat ion of the l i fe course, as the durat ion,
sequencing, and d i rec t iona l i ty of movement be tween his tor ica l ly age-
gr ad ed s tatus posi t ions are bec om ing increasingly diverse and subject to
m anifo ld co ntingencies (see O Ra nd, this vo lum e; M odell , 1989). Th ese
t ren ds m ak e the l inkages be tw een school an d wo rk, as wel l as occ upat ion a l
career l ines fol lowing completion of formal education, increasingly tenu
ous .
U nd er such c ircumstances , i t m ig ht be a rg ued tha t soc ia l psychologica l
fac tors become even more important in de termining the achievement
pro cess , no t only in the early s tages of the life course, bu t th ro ug ho ut the
soc io-econom ic career . N o longer can it be assum ed tha t indiv idua ls , once
securely placed on a lower rung of a success escalator, can ride smoothly
tow ard the top , carr ied forward s t ruc tura l ly v ia s t rong ins t i tu t iona l ized
career tracks . W ith a wid er a rray of bran ching po ints , m ore prev a lent
reversa ls in the d i rec t iona l i ty of m ove m ent , and m ov em en ts be tw een such
tracks (e .g. , re tu rn to school for con tinu ing trainin g), ind ivid ua l m otiva tion,
voli t ion, an d effort m ay beco m e increasingly sal ient . T hu s, w he rea s re-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 46/375
9
search on the social psychology of mobili ty has been highly focused on
adolescents and youth, future investigators might f ruitfully extend the
scope of s ta tus a t t a inm ent s tud ies to o lde r people , exam in ing h um an
agency among middle-aged and older workers as de terminants of career
atta inm en t later in the life cou rse.
In light of the inc reas ing dive rsification of career seq ue nc es, it is per t inen t
to investigate the manner in which success aspirations inf luence attain
m en ts as careers unfold . Clause n (1993) exam ines the a t ti tudinal prec ursors
in adole scence of l ife-long attain m ent, f inding com peten ce to be especially
pred ictive . But are there any dis t inct social psych ologica l antec ed en ts of
atta inm en t for pe rso ns of different age? A teenager , upo n leaving full- t ime
scho oling, mi gh t be mo tiva ted to f ind a job, at least in par t , by a desire for
independence and dissa t is fac t ion with cont inued economic and other
forms of dependency upon parents (Borman, 1991) . In contrast , the older
w ork er ' s job search and re location a t tem pts , an d other achievem ent- re la ted
behaviors , may be more closely l inked to family economic needs, the
po sitio n of w or k in the hie rarc hy of ide nti ties (Stryker, 1985) an d the level
of inves tme nt in , an d com m itme nt to , the work role . W hat are the psych o
logical an tece de nts of fur ther m obili ty on the pa r t of those wh o are m id-w ay
on career ladde rs? Are there psychologica l d im ens io ns which dis t ing uish
those w ho pl ate au relatively ear ly in their careers , f rom those w h o
cont inue to progress?
The re ha s be en relat ively li t tle s tu dy of the social psychological dy na m
ics of achievement in middle age , when expected a t ta inments do not
mater ia l ize , or w he n unan t ic ipated obs tac les are enco unte red. W hat psy
chological factors promote ear ly recovery from job loss and rapid re
em plo ym ent? Ha m il ton, e t a l .' s quas i -exp er imenta l s tudy of w ork ers in
four closing (and a control group of twelve non-closing) General Motors
pla nts f rom 1987 to 1989 found that dep ressio n w as a s ignif icant pre cu rsor
of sub sequ ent inabi li ty to locate a new job. Both un em plo ym en t an d
depress ion a t w av e pred ic ted unemp loym ent a t wav e 2 , one ye ar after the
plant c losed; and unemployment and depress ion a t wave 2 s imilar ly
predic ted unemployment two years af ter the c los ing.
H ow ev er , K essler , et al (1989), on the basis of their s tud y of a co m m un ity
sam ple in areas of Detroi t wi th high un em plo ym en t
ra tes ,
repo rt that greater
pr ior d is t ress pred ic ted reem ploy m ent of the un em plo ye d ov er a one-year
per io d (con troll ing age, sex, edu catio n, race, and m arital s tatus) . The
au tho r s specu la te tha t the mo re d i s t re s sed unem ploy ed w orker s may hav e
eng aged in m ore s t renu ou s job search, enco urage d by im pro vin g econ om ic
condi t ions .
Preexist ing psychological or ientations may inf luence the person's def i
nit io n of the s i tuatio n up o n loss of a
job.
Un emp loyed worker s may de fine
themselves as unemployed members of the labor force and ac t ive ly seek
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 47/375
3
reemployment. Alternatively, they may view themselves as temporarily
out of the workforce, and p ostpone seeking w o rk - a more typical definition
am ong w orkers in closing General Motors plants w ho had mo re seniority,
were older, and female (Ham ilton, et
al.,
1991). Many unemp loyed wom en
w ithdraw from the labor mark et
to
pursu e family-related objectives (W arr,
et al.,
1982).
If the recent jobless see little prosp ect of finding a new job, they
ma y become discouraged w orkers and drop out of the labor force
entirely. Finally, if prosp ects are not brigh t, and if jobless workers are old
enough, they may define themselves as retired.
It is reasonable to supp ose that those with lower self-efficacy would be
more likely to become discourag ed workers (Banks and ah (1988:118).
Higher levels of depression and self-blame may also predict withdrawal
from the labor force upo n losing a job (Hoffman, et al, 1991: Table 6). In fact,
not looking for a job may be one way to protect a fragile psych e, for Hoffman
et al. (1991) show that to seek a job and not to find one increases dep ressio n.
Jobless workers in their study wh o were not looking and did not find
work were less distressed than the active but unsuccessful job seekers.
Paradoxically, given the various possible definitions of the jobless situation,
those wh o become unemployed could have high er self-efficacy than those
w ho become discouraged or retired.
Some have speculated that differences in job search behav ior mediates
the linkages between psychological variables and em ployme nt outcomes
(Feather, 1986; Kessler, et al., 1989). But there has been little system atic
stud y of this possible source of mediation. One study of une mployed Israeli
worke rs (Shamir, 1986) found that among those who becam e reem ployed,
those with higher self-esteem made more use of individualistic, informal
and active methods of job search, such as the use of personal contacts and
direct application to the prospective employer. Those with lower levels of
self-esteem were more likely to use im person al metho ds, such as the labor
exchange or employment services. They were also more willing to accept
jobs that did not match their income and job content goals. W hereas such
flexibility m ay be useful with respect to imm ediate em ployment prospects,
it may have more negative long-term career implications.
The Causal Priority of Structure
and chievement Orientations
Some may question the causative role of achievement orientations in
attainment, arguing that attitudes reflect likely career destinations, given
socio-economic realities and existing opportunities, rather than determin
ing the direction or extent of actual achievem ent (Roberts, 1968). In
em phasizing socialization as the link between social origin and destination ,
the status attainm ent researchers ten ded to assum e, implicitly if not explic-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 48/375
itly, tha t
the
individual
is
relatively free
to
move within
the
social structure;
social structural constraints on the attainment process were relatively
ignored (see Kerckhoff, 1976,1984). More recent work, such as those cross-
national studies referenced above, em phas izes structural allocation, view
ing attainm ent as resulting from structura l limitations and selection criteria,
and the ind ividu al as relatively constrained by the social struc ture. If
oppo rtunities determ ine achievement orientations (Ogbu, 1989; Kerckhoff,
1994),
these psychological variables might merely mediate the effects of
structural forces.
It is not easy to decide among such causes, for the identification of
antecedents, correlates, or predictors will dep end on the particular point at
which one enters the causal sequence. Taking und erclass youth as a case
in point, we know much about the economic, technological, demographic
and other structural trends that have eroded youth employment opportu
nities.
Youth in the inner
cities
often have few occupational
role
models and
very limited job prospects. However, they do have chances to obtain
income through hustling, crime, drug s, or welfare dependency. Educa
tional and occupational aspirations, the motivation to seek jobs, and the
sense of control over attainment are undoubtedly affected (Wilson, 1987;
Sullivan, 1989; Banks and Ullah, 1988). Unsuccessful a ttempts to find paid
work, accompanied by perceived or
real
discrimination, may only intensify
pessimism and weaken the propensity to work. Furnham (1985:109)
speaks of a des tructive vicious circle occurring when you th fail to find
jobs: they experience stress and disappointment, lowered self-esteem and
decline in expecta tions , all of which lessen the intensity and fruitfulness of
subsequ ent job search. In view of this sequence, is it the psychological
impedimen t, or the structure, which is at the root of the failure to succeed?
This simple formulation of the question does not do justice to the complex,
dynamic, interactive, and reciprocal interrelations of structural, psycho
logical and behavioral phenomena in the achievement process. Future
researchers need to attend to both struc tural and psychological variability
to
fully unders tand
the
processes of achievement throughout ind ividual lives.
References
Alexander, Karl L., Bruce
K.
Eckland, and Larry J. Griffin. 1975. The Wisconsin
Model of Socioeconomic Achievement: A Replication.
merican Journal of
Sociology
52:324-342.
Ashton, David N. and Johnny Sung.
1991.
Labor Markets and the Life Course
Patterns
of
Young Adults in
Great Britain.
Pp. 23 42 in The Life Course and Social
Change
edited by
Walter.
R. Heinz. Wannheim: Deutscher Studien Verlag.
Baker, Meredith and Peter
Elias. 1991.
Youth Unemployment and Work Histories.
Pp.
214 244 in Life and Work History Analyses: Qualitative and Quantitative
ev
opments
edited by Shirley Dex. London: Routledge.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 49/375
32
Ba ndu ra, Albert . 1988. Org anizat ion al Applicat ions of Social Co gnit ive Th eory .
Australian
Journal
of Management
13:275-302.
Banks, M. H. an d P . Ullah . 1988. Youth Unem ployment in the 1980 s: Its
Psychological
Effects. London : Croom He lm.
Beck, E. M., Patrick M. Ho ra n, and Ch arle s M. Tolbert II. 1978. Stratification in a
Dual Economy: A Sectoral Model of Earnings Determinat ion. American Socio
logical
Review 43:704-720.
Borman, Kathryn M.
1991.
The
First Real Job:
A Study of Young Workers. A lbany : State
Universi ty of New York Press.
Cal l , K ather ineT ., Jeylan
T
Mort imer , Ch aim un Lee, and Kather ine D enneh y . 1993 .
H igh Risk You th and the Atta inm ent Process. Pap er pres ented at the annu al
m eet in gs of the Am erican Sociological A ssociat ion.
Clausen, John S. 1993. Am erican L ives. Looking Back at the Children of the Great
Depression. New York: Free Press.
Coe, Richard D. 1978. Ab senteeis m from W ork. Ch. 5 in Five Thousand American
Families
—
Patterns of Econom ic Progress, Vol. VI, Accounting for Race and Sex
Differences in Earnings and O ther Analyses of the First Nine Years of the Panel Study
of Income Dynamics, ed i ted by Greg J . Duncan & James N. Morgan . Survey
Research Center, Institute for Social Research: University of Michigan.
Corcoran , M ary and Martha S. Hi l l. 1980. Pers i s tence In U nem ploy m ent A m on g
Ad ul t M en. Ch. 2 in Five Thousand American Families — Patterns of Economic
Progress, Vol. VIII Analyses of the First Eleven Years of
the
Panel Study of Income
Dynamics, ed i ted by Greg J. Du ncan & James N. Morgan . Surve y R esearch
Center, Institute for Social Research: University of Michigan.
D avis , Jam es A. 1964.
Great
Aspirations:
The Graduate School Plans
of Am erica s
ollege
Seniors. Chicago: Aldine .
D enn ehy , Ka therine and Jeylan T. M ort imer. 1993. W ork and Fam ily Orie ntat io ns
of Co nte m por ary Ad olescent Boys and G irls . Pp . 87-107 in Men, Work and
Family, edi ted by Jane C. H ood . N ew bu ry Pa rk, CA: Sage Publicat ions.
DiPre t e, Thom as . 1981. U nem ploy m ent over the Life Cycle: Racial Differences and
the Effect of Changing Economic Condit ions. American Journal of Sociology
87:286-307.
Duncan, Otis D., Daniel L. Featherman and Beverly Duncan. 1972. Socioeconomic
Background and Achievement. New York: Seminar Press.
Fea the r, Nor ma n
T
1986. Em ploym ent Imp ortance and Help lessness abo ut Poten
t ia l U nem ploym ent am ong Studen ts in Secondary Schools. Australian
Journal of
Psychology 38:33-44.
Feather, N or m an T. and Go rdo n E. O'Brien. 1986. A Lo ngitudina l Stu dy of the
Effec ts o f Em ploym ent and Un em ploym ent on School -Leavers .
Journal
of Occu
pational Psychology
59:121-144.
Fe athe rm an, D avid L. and Ken neth I. Spen ner. 1988. Class and the Socializat ion of
Ch ildren; Con stancy, Ch ang e, or Irrelevan ce? Pp. 67-90 in Child Development in
Life Span Perspective, edi ted b y E. M avis He thering ton, R ichard M. Lerner, a nd
M arion Ferm utter. Hil lsdale , NJ: Law rence Erlb aum Associates.
Finch, Michael D., Michael Shanahan, Jeylan T. Mort imer, and Seongryeol Ryu.
1991. W ork Experience and Con trol Or ientat ion in Ad olescen ce. American
Sociological
Review 56:597-611.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 50/375
Fre em an, Richard B. and David A . W ise. 1979. Youth Unemployment. C a m b r i d g e ,
Massachuset ts: Nat ional Bureau of Economic Research.
Fu rnh am , A drian . 1985. Youth Unem ploym ent : A Review of the Li te ra ture .
Journa l of Adolescence
3:109-124.
G eca s, Viktor. 1979. Th e Influence of Social Class on Soc ialization, Pp . 365-404 in
Contemporary T heories about the Family Vol. I Research Based Theories
edi ted by
W esley R. Burr, Re ub en H ill , F. Ivan N ye , an d Ira L. Reiss. NY : Free Pre ss.
G eca s, V iktor . 1989. Th e Social Ps ych olo gy of Self-Efficacy.
Annual Review of
Sociology 15:291-316.
Ge issler, B. an d Helg a Krug er. 1992. Balan cing the Life C ou rse in Re spo nse to
Inst i tut ional Requ irem ents . Pp . 151-167 in
Institutions and Gatekeeping in the Life
Course edi ted by W al ter Heinz . Wann heim: D eutscher Stud ien Verlag .
G ord on , Ch ad. 1972. Loo king A head : Self-Conceptions, Race, and Fam ily as
Determinants o f Adolescen t Orien ta t ion to Achievement .
Rose Monograph Se-
ries.
W ash ingto n, D C: Am erican Sociological A ssociat ion.
Ha gestad , Gu nhi ld . 1992 . Ass ign ing Rights and Dut ies : Age , Du ra t ion , and
G en de r in Social Inst i tut io ns. Pp . 261-279
in Institutions and Gatekeeping in tltehife
Course edi ted by Wal ter Heinz . Wannheim: Deutscher Stud ien Verlag .
Hamil ton, Stephen F. 1990. Apprenticeship for
Adulthood
Preparing Youth for the
Future. New York: The Free Press.
H am il ton, V. L., Hoffman, W. S., Broman, C. L. & Raum a, D. 1991. Afterm ath: A
Pane l S tudy of Un em ploym en t and Men tal Hea l th am ong Au tow orke rs . Un
publ i shed paper .
Hau ser , Rober t M. 1971.
Socioeconomic Background
and Educational Performance. Rose
Monograph Series.
W ashin gton , DC: Am erican Sociological A ssociat ion.
Hoffm an, W . S., V. L. H am il ton, C. L. Brom an, and D. Ra um a. 1991. U nem plo y
me nt , Dep ress ion , an d Self-Blame A m on g Au tow orke rs . Pap er p rese n ted a t the
annual meet ings of the American Sociological Associat ion Annual Meeting.
Hu i , W eng-ta t. 1991 . Reserva t ion W age Analys is of U nem ploye d Youths in
Aust ra l ia .
Applied Economics
23:1341-1350.
Jones,
Step hen R. G. 1989. Reserv at ion W ages and the Cost of Une m plo ym en t .
Economica
56:225-246.
erckhoff Alan C. 1972. Socialization an d Social Cla ss. En glew ood Cliffs, NJ:
Prent ice Hall
erckhoff Alan C. 1974. A mb it ion and A ttainm ent: A Study of Four Sam ples of
Am erican Boys. Rose Monograph Series. W ash ing ton , D C: Am erican Sociological
Associat ion.
erckhoff Alan C. 1976. The Status A ttainm ent P rocess: Social ization or Alloca
t ion?
Social Forces
55:368-381.
erckhoff A lan C. 1984. Th e C ur ren t State of Social Mo bility Re sea rch. The
Sociological Quarterly
25:139-153.
erckhoff A lan C. 1989. O n the Social Psy cholo gy of Social M obility Proc esse s.
Social Forces
68:17-25.
erckhoff Alan C. 1990. Getting Started Transition to Adulthoo d in Great Britain.
Boulder: W estview Press.
erckhoff Alan C. 1993. Diverging Pathways.
Social
Structure and
Career
Deflections.
Ne w York: Ca m brid ge U nivers i ty Press .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 51/375
34
Kerckhoff Alan C. 1994. Social Stratification a nd M obility Proce sses: Th e Interac
tion Betw een Ind ivid ual s an d Social Str uct ure s. Pp . 476-496 in
So ciological
Perspectives on Social Psychology, edi ted by Karen Cook, Gary A. Fine, an d Jam es
S. House. New York: Allyn and Bacon.
Kessler, Ronald C , J. Blake Turn er, and James S. H ouse . 1989. U ne m plo ym en t ,
Reem ploym ent , and Emot ional Funct ion ing in a Co m m uni ty S am ple .
American
Sociological
Review
54:648-657.
Kohn, Melvin L. 1969. lass and Conformity: A Study in Values. Hom ew ood , IL.:
Dorsey .
Ko hn, Melvin L. 1981. Pe rson ality, O ccu patio n, and Social Stratification: A Fram e
of Refere nce. Pp . 267-297 in Research
in
Social
Stratification and M obility,
edi ted by
Do nald J . Treim an an d Robert V. Robinson . Green wich , CT: JAI Press, Inc.
Kohn, Melvin L. & Carmi Schooler. 1983.
Work and Personality: An Inquiry into the
Impact of Social Stratification, N orw ood, N ew Jersey : Ablex .
Lynch, Lisa M. 1983. Job Search and Y outh U ne m plo ym ent . Oxford Economic
Papers 35(Supplement):27\-282.
M cLaug hlin, Steven D ., Barbara D . Melber, John O.G. Billy , Den ise M. Zim m erle,
Linda D. W inges, and T erry R. Johns on. 1988. The Changing Lives of American
Women.
Ch apel Hil l , N C: The Un iversi ty of N or th Carol ina Press.
M cN all , Miles, Tim m othy Du nnig an, an d Jeylan T. M ort im er. 1994. The Educ a
t ional Ac hieve m ent of the St. P a u l H m o n g . Anthropology and Education Quarterly
25:1-22.
Maccoby, Eleanor E. and Carol N. Jacklin. 1974.
The Psychology of Sex Differences.
Stanford: Stanford Universi ty Press.
M eyer, R. M. & David A . W ise. 1982. H igh School Pre para t ion an d E arly Labor
Force Ex perie nce. Pp . 277-347 in
The Youth Labor Market Problem: Its Nature,
Causes, and Consequences,
edi ted b y Richard B. Free ma n an d D avid A. Wise.
Chicago : Universi ty of Chicag o Press.
M il lham , Spen se, Roger Bullock and K enne th Ho sie. 1978. Juveni le U nem ploy
ment: A Concept Due for Re-Cycling? Journal of Adolescence 1:11-24.
Mode l l j ohn . 1989 . Into One s Ow n. From YouthtoAdtdthoodintheUnitedStatesl920-
1975.
Berkeley an d Los An geles, CA : Un iversi ty of Cal ifornia Press.
M ort im er, Jeylan T. 1994. Ind ividu al Differences as Prec urso rs of Yo uth Un em
ploy m ent . Pp . 172-198 in
Youth Unemployment and Society,
ed i ted by Anne C.
Petersen . N ew York: Ca m bridge Univers i ty Press.
M ort imer , Jey lan
T.
and M ichael D. Finch. 1986. Th e Effects of Pa rt-T im e Wo rk on
Self-Concept and Achiev em ent . Pp . 66-89 in
Becoming a Worker,
edi ted by
Kathryn Borman and Jane Reisman. N orw ood , New Jersey: Ablex .
M ort im er, Jeylan T. and Do nald K um ka. 1982. A Further Exam inat ion of the
'Occupat ional Linkage Hypothes is . '
The
Sociological
Quarterly
23:3-16.
M ort im er, Jeylan T., Jon Lorence an d D onald Ku m ka. 1986.
Work
Family, and
Personality: Transition to Adidthood. Norw ood, Ne w Jersey : Ablex .
M ort im er, Jey Ian T. Seongry eol Ryu, Katherine D enn ehy , El len Efron Pim entel , an d
Ch aim un Lee . 1994. Par t -Time Work and Occupat ional Value Form at ion in
Ad olescenc e. Revision of pap er pres ented at the 1992 A m erica n Sociological
Associat ion Annual Meeting. Pi t tsburgh.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 52/375
35
O'B rien, G ord on E. an d N or m an T. Feather. 1990. The Relative Effects of Une m
ployment and the Qual i ty of Employment on the Affec t , Work Values and
Personal Con t ro l of Adolescen ts .
journal of Occupational Psychology
63:151-165.
O gbu, John U. 1989. Cul tura l Boundaries and Minori ty Youth Orien ta t ion T ow ard
W ork Pre par at io n. Pp . 101-140 in Ad olescence and Wo rk Influences of Social
Structure Labor Markets and Culture edi ted by David Stern and Do ro thy Eichorn .
Hil lsdale, New Jersey: Lawrence Erlbaum.
O ster m an, Paul . 1989. The Job M arket for Ad olesce nts. P p. 235-256 in dolescence
and Work Influences of Social Structure Labor Markets and Culture edi ted by David
Stern & D oro thy Eichorn. Hil lsdale, Ne w Jersey: Law rence E rlbaum ,
Rindfuss, Richard R., Elizabeth C. Cooksey and R. L. Sutterlin. 1990.
Young Adult
Occupational Achievement: Early Expectations Versus
Behavioral
Rea lity.
Pape r p re
sented at the World Congress of Sociology. Madrid, July.
Robert s , K. 1968. The Ent ry in to Em ploym ent : An A pproa ch To w ard s a Genera l
Theo ry . Sociological
Review
16:165-84.
Rosenbaum, James E., Takehiko Kariya, Rick Set tersten and Tony Maier. 1990.
Market and Network Theories of the Transi t ion from High School to Work:
Their Ap plicat ion to Indu strial ized Societ ies. Annual Review of Sociology 16:263-
299.
Rosenberg, Morris . 1957.
Occupations and Values.
Glen coe, Il l inois: Free P res s.
Rutter, Michael . 1986. The D evelo pm ental Psych opath ology of De pressio n: Issues
an d Per spe ct ives . Pp . 3-30 in
Depression in Young People: Developmental and
Clinical Perspectives edited by Michael Rutter, Carroll E. Izard & Peter B. Read.
N ew York: Guilford Press.
Schw albe, Michael L. and Viktor Gecas. 1988. Social Psychological D im ensio ns of
Job-R elated Disability. Pp . 233-271 in Work Experience and Psychological Develop-
ment Through the Life Span edi ted by Jeylan T. Mort imer & Kathryn M. Borman.
Boulder: Westv iew Press .
Sewell , W il l iam H., Ar chibald O. Haller, an d Alejandro P ortes. 1969. The Edu ca
t iona l and Ear ly Occupat ional A t ta inment P rocess . American Sociological Review
34:82-93.
Sewell , Wil l iam H., Archibald O. Haller, and George W.
Ohlendorf
1970. The
Edu cat ional and Early Oc cupa t ional Status A ttainm ent Process: Replication an d
Revision. American Sociological Review 35:1014-1027'.
Sham ir, Boaz. 1986, Self-Esteem and the Psychological Imp act of Un em plo ym en t .
Social Psychology Quarterly
49:61-72.
Sh ana han , Michael J., Michael D. Finch, Jeylan T. M ort ime r, an d Se ongryeo l Ryu .
1991. A dolesc ent W ork Experience an d De pressive Affect. Social Psychology
Quarterly
54:299-317.
Simmons, Roberta and Dale Blyth. 1987. Moving into Adolescence: The Impact of
Pubertal Change and School Context.
New York: Aldine.
Spe nner, K enne th I. and Lu ther B. Otto. 1985. W ork an d Self-Concept: Select ion
an d Socialization in the Early Ca ree r. Pp . 197-235 in Research in Sociology of
Edu cation and Socialization Vol. 5 edi ted by Alan C. Kerckhoff Greenwich , CT:
JAI Press.
Spenne r , Kenn eth I. and David L. Fea therm an. 1978 . Ach ieveme nt A m bi t ion s .
Annual Review of Sociology 4:373-420.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 53/375
36
Stryker, Sheldon. 1985. Symbolic Interaction and Role Theory. Pp . 311-378 in
Handbook o f Social
Psychology Vol I, edited by Gardner Lindzey & E. Aronson.
New York: Random House.
Sullivan, M ercer L.
1989. Getting
Paid:
Youth Crime and Work in the Inner
City. Ithaca:
Cornell U niversity Press.
W arr, Peter, P. Jackson, and M. Banks. 1982. Duration of Unemploym ent and
Psychological Weil-Being in Young Men and Women. Current Psychological
Research
2:207-214.
W ilson, Kenneth L. & Alejandro Portes. 1975. The Educational A ttainm ent Process:
Results from a National Sam ple.
American journal o f Sociology
51:343-363.
Wilson, William
J.
1987. The Truly
Disadvantaged:
The Inner City tlie
Underclass
and
Public
Policy. Chicago: University of Chicago Press.
Winefield, Anthony H ., and M arika Tiggemann.
1985.
Psychological Correlates
of
Employment and Unemployment: Effects, Predisposing Factors, and Sex Differ
ences. Journal o f Occupational Psychology
58:229-242.
William
T.
Grant Foundation. Commission on
Work
Family, and Citizenship . 1988.
77K
Forgotten
Half
Pathways to Success for America s Youth and Young Families.
Washington, D.C.: The William
T.
Grant Foundation.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 54/375
Building Conceptual
and Empirical Bridges
be tween Studies of E ducational
and Labor Force Careers
lan C
erckhojf
uke University
Th ere tend to be tw o gro up s of resear chers w ho d eal with social s trat if i-
ca t ion processes , one group working on educat ional processes , the o ther
g ro up wo rk ing on labor force processes . Th ere is only very l im ited
com mu nica t ion be tw een those two g rou ps , and w e a re far from an adeq ua te
un de rs ta nd in g of the com bine d effects of those two s trat ificat ion proc esses
and h ow they are re la ted to each other .
If w e consid er the bodies of l i terature tho se two g ro up s of scholars h av e
p ro d u ce d , it is po ssib le to identify at least six ne ste d levels of ana lysis of th e
factors mat contribute to the dis tr ibution of a cohort of individuals into
locations in a so ciety s s tratif icat ion system :
1
individual at tr ibutes (e .g. ,
am bit ion , abil i ty) , inter per son al relat ion ships (e .g., family a nd pee r gr ou p
resources , va lue def in i t ions , encouragement) , educat ional se t t ings (e .g . ,
in terna l school organiza t ion, curr iculum , teaching
style),
occup ation s (e .g. ,
ski ll level , au ton om y) , w or k set t ing s (e .g. , f i rm size, intern al labo r m ark ets ,
service industries), nation states (e.g. , socialist, corporatist, capitalist).
I t is no t possible to deal with al l s ix in the pre sen t discussio n, ho w ev er.
I wi l l focus on the las t four—educat ional se t t ings , occupat ions , work
se t t ings , and na t ion s ta tes in order to develop some genera l ideas about
stratification p roc ess es as the y var y across societies. I w ill focus on th e
social ly p rov ide d c han nels of at tain m en t as a m ea ns of identifying societal
varia t ion. In do in g so, I wil l sugg est the outl in es of a m ap of s tratif ication
processes that may be useful in further comparative s tudies .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 55/375
8
The educa t iona l sys tem has app ropr ia te ly been re ferred to as the
sort in g mac hin e for society (Spring 1976), bu t I w an t to br oa de n that
metaphor to inc lude both educat iona l and labor force sor t ing processes .
This broader metaphor wil l help to make two features of s trat if icat ion
pro ces ses mo re salient. O ne is the critical significance of the socially
provided s t ruc tura l loca t ions whose h ierarchica l order cons t i tu tes the
sub stan ce of w hat w e m ean by s trat if ied. The other
s
th e fact tha t socie tal
inst i tut ion s act ively contrib ute to dis tr ibu ting ( sort ing ) a coh ort into
thos e s trat if ied locations. This is no t to m inim ize the im po rtan ce of the role
played by individual actors and their int imate social re lat ions (which
Mortimer and Bidwell e t a l demonstrate in their contributions to this
vo lum e), bu t only to m ak e sal ient the act ive role of inst i tut ional actors .
If we conceive of s trat if icat ion processes as those which generate the
dis tr ib utio n of a coho rt into hierarchical ly arra ng ed pos it ions in the society,
give n their origin s in that sam e set of posi t ion s , w e see that th ose s trat if ica
t ion processes sha pe the life course careers of the cohor t m em be rs . Th at is ,
the sort ing pro cesses serve to shap e the origin -dest in ation trajectories of the
cohort members , and the ins t i tu t iona l ly provided s t ruc tura l loca t ions de
fine th e alter na tive trajectories that are po ssib le. Th ose trajectories ar e pa th s
th ro ug h sets of s truc tura l locations enc oun tered at successive s tages of the
life cou rse, an d w hich ind ivid ua ls fol low w hich trajectories is de term ine d
by the intersection of individual and inst i tut ional act ions.
Th e very c on cep ts of caree rs an d trajectories, if they are to be at all u seful
in social science research, must be based on the postulate that there are
varying probabil i t ies of movement from a given location at one s tage to
severa l possible locations at a successive s tage. That is , it sho uld b e poss ib le
to identify care er
lines,
or mo s t f requently t rave led pa thw ay s throug h the
multiple locations across life course stages (Gaertner 1980; Spilerman 1977;
Sp enn er, O tto, and Call 1982).
Our research has tended to separate the analysis of careers and career
lines in schools from th e ana lysis of those in the labor force. W e ne ed to
br idge those two separa te approaches by conceptua l iz ing s t ra t i f ica t ion
processes as occurr ing throughout the l i fe course (as O 'Rand sugges ts in
this volum e). To do so m ea ns to view ca reers and career l ines as l ife time
trajectories which result from strat if icat ion processes occurring in both
educational and labor force set t ings . I t a lso means that we seek to under
sta nd the pro bab ility of linka ges be tw ee n locations (steps in trajectories) n ot
just w ithin schools and w ithin the labor force bu t a lso be tw een th em . W e
can use th e con cep ts of career, career line an d trajectory as tools to g en era te
a map of s trat if icat ion processes throughout the l i fe course.
I wil l beg in by focusing on the s tructu ral features of edu catio nal system s
an d ho w the y ma y affect the proces s of social stratification. I will use a
comparative perspective to point up the peculiar features of the American
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 56/375
39
educational system and how those features may be related to s trat if ication
pro ces ses. I will th en t ur n to som e structu ral features of the labor forces of
indus tr ia l socie t ies , again wi th an emphas is on how the Amer ican labor
force structures differ from those in other industrial societies. And, finally,
I wil l spec ulate a bit abou t w ays in wh ich these features of edu catio nal an d
labor force s tructures may be related to each other and how their var ied
association s across societies m ay he lp to pro du ce var ia t ions in s trat ification
processes and in the trajectories cohorts follow in different societies.
ducational Structures and Stratification Processes
Som e recent co m pa rativ e wo rk ha s focused o n differences in the ed uc a
tiona l sy ste m s of indu stri al societies an d on the effects of tho se differences
on the dis tr ibution of individuals in the societies ' s trat if ication systems
(A llm end ing er 1989; M aurice , Sell ier , an d Silvestre 1986; M uller an d Karle
1992; Raffe and To me s 1987). A ltho ug h all edu catio nal syste m s sor t
s tu de nt s into hierarchical catego ries , they differ in the w ay s they go ab ou t
do ing i t an d in the natu re of the hierarchies they pro du ce .
The edu catio nal system s vary in several im po rtan t way s: (1) the de gre e
of stratif ication of the sys tem (i.e. , the pro po rtio ns of a co ho rt that are able
to m ov e to successively h igh er a t tain m ent levels) , (2) the deg ree of central
ized con trol of the system an d the uniform ity of i ts pro gra m s and pr od uc ts ,
(3) h ow differentiated the educa tion al creden tials are, (4) the e du catio nal
sy ste m 's cap acity to str uc tu re the flow of stu de nt s int o the labor force (i.e. ,
the match b etw een educ at ional credent ia ls and occupat ional ca tegor ies), (5)
the degree a nd k ind of ins t i tu t ional l inkage betwee n ed ucat ion al and labor
force insti tuti on s - there m ay be very little linkag e (as in Franc e), the re m ay
be a direct l inkage (as in Jap an [Rose nbau m, K ariya, Setters ten, an d M aier
1990;
Spile rma n a nd Ishida in this volume ]) , or the l inka ge ma y be du e to a
credent ia l ing sys tem that depends heavi ly on apprent iceships and other
types of labor force experience (as in Germany [DiPrete and McManus in
this volume]) .
The U.S. edu cat ion al sys tem is essentia lly un ique am ong indu s tr ia l
societies. In terms of the five dimensions of variation just noted, it differs
from m os t sys tem s in oth er ind us trial societies in the following w ay s: (1) It
is much less s trat if ied; more American s tudents s tay in school at higher
atta inm en t levels tha n do s tud en ts elsewhe re. (2) I t is hig hly dec entraliz ed,
an d there is
grea t dea l of var iat io n am on g both the pr og ram s offered at any
par t icular edu cat ion al level and the qua l i ty of the s tud ents w ho com plete
thos e pr og ram s. (3) Th e A me rican system offers only very gen eral cred en
tials (a high school diploma or a college degree) while most European
sy ste m s offer a m u ch w id er rang e of differentiated cre de ntia ls. (4) As a
resul t , mo st Eu rope an sys tems hav e a m uc h grea ter capaci ty to s t ruc ture
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 57/375
4
the flow of students into the labor force; there is a tighter fit between
educational credentials and occupational categories in Europe than in the
U.S. (5) There is less formal l inkage betw een sch ools and em plo ye rs in the
U.S.
than in almost any other industr ial society. Students generally leave
school and seek jobs on their own.
The edu catio nal system s of all ind ustr ia l societies use s tructu ral dis t inc
t ions to orga nize their s tu de nts into grou ps or un its receiving differentiated
curr icu la. In m os t Eu rop ean societies specialized secon dary schoo ls are
used to accomplish this purpose (e .g. , the Hauptschule, Realschule, and
G ym na siu m in Ge rman y) , bu t there m ay a lso be di f ferentia ted pro gra m s
w ithin the schools . Am erican schoo ls are m uc h less l ikely to be differenti
ated in this way , al th ou gh the increasing use of m ag ne t schools is a m ov e in
that direction, and organizational units within schools (abil i ty groups,
tracks) are generally found .
Curr icular d is t inc t ions wi thin schools are even more impor tant in the
stratif ication proces s in the U.S. than in mos t other indu str ial societies just
because there are few other kinds of s trat ifying mechanisms used by the
edu catio na l system . S tu de nt s ' locations in those differentiated s truc tura l
posit ions are necessar i ly based on insti tutional decis ions, if for no other
reason because of organizational constraints (Hall inan 1992) . The decen
tralized (local) con trol of schoo ls in this co un try leaves that de cisi on -m ak ing
process more open to what Bidwell and Quiroz (1991) refer to as "client
power" (parental intervention) than are s imilar decis ions in societies with
m ore cen tralized co ntrol of the schools . I t seem s likely, then , that such
decisions will reflect the social status of students ' parents (Lareau 1989;
U seem 1992; W illms in this vo lum e). In any ev ent, the internal orga niza
t ional dis t inctions serve to increase the spre ad of acad em ic perfo rm an ces of
s tud ents as G am oran a nd A lexander and E ntwis le discuss in the i r contr ibu
tions to this volume.
A ltho ug h they hav e received less system atic at tentio n in the l i teratu re on
the s trat ifying effects of edu cation al inst i tutio ns, var iat ions in the ac ade m ic
programs in pos t - secondary ins t i tu t ions undoubtedly serve to fur ther
increase that spre ad. These var iat ion s are of tw o kin ds, am on g kin ds of
insti tu tions ( technical inst i tutes , co m m un ity a nd junior colleges, four-year
colleges, universi t ies) and a m on g the "qua li ty" levels wit hin each of these
(Ivy Leag ue, small s tate unive rsi ty) .
Th ese character is t ics of ou r edu catio nal syste m lead it to pr od uc e a set of
"g ra d ua te s" (of either high school or college) w ho are highly v ar ied in their
edu catio nal quali t ies . Yet, these high ly var ied "p ro du cts " of ou r school
syste m are relatively undifferen tiated in term s of the formal c red en tials
they obtain.
2
Th ere are four large am or ph ou s categories of pro du cts of the
Amer ican educat ional sys tem—high school drop-outs , h igh school gradu
ates,
col lege drop-outs (who far outnumber high school drop-outs ) and
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 58/375
4
col lege graduates .
If we think of this outcome in terms of career trajectories, it is apparent
that a nu m b er of relat ively refined location dis t inct ions with in the ed uc a
t ional syste m (defined in term s of abil ity gr ou ps, tracks and so on) seem to
lead to a ve ry delim ited set of term ina l locations. Ed uca tiona l careers
may be t raceable through s t ruc tura l loca t ions in e lementary school , be
tween e lementary and secondary school , and be tween secondary and
po st-sec on dar y schools , bu t the end s tates are no t formally highly differen
t ia ted.
The Education Occupation Linkage
When we fo l low American s tudents ' careers a f te r they comple te the i r
schooling, i t is a lso apparent that the credentials they obtain are seldom
directly l inked to kin ds of
jobs.
There a re few obvious educ at ion-occ upat ion
connect ions . Ove ra l l , there i s a very loose coupl ing be tw een Am erican
educational and labor force inst i tut ions, a l though i t may be possible to
ident i fy l inkages be tween the very genera l educa t iona l c redent ia ls and
fairly gro ss occu patio nal categories (e.g. , hig h school dr op -o uts an d college
gra du ates a re not very of ten foun d in the same o ccupat ions) .
This very loose educat ion -occup at ion l inkage in the U.S. is u n d o u b t e d l y
one of the reasons for the divis ion of labor between those s tudying educa
t ion an d those s tud yin g the labor force. For the majori ty of A m erica n
s tud en ts , the educat io na l sys tem p rov ides l i tt le , if any, d i rec t pre pa ra t io n
for labor force par t ic ipa t ion . Most Am erican s tud en ts leave school w i th
neith er occu pation ally-relev ant prep ara t ion no r cred entials . Specific job
skills are learn ed after lea vin g school. O r at least after leav ing the only kin d
of school Am ericans see as w orth count ing .
3
W ork skills are lear ned
either on-the-job or through part- t ime courses at technical inst i tutes or
co m m un i ty col leges . I t is un us ua l for e i ther on- the- job t ra in in g or
pos t -se cond ary v oca t iona l t ra in ing courses to lead to na t iona l ly recognized
occupat iona l ly- l inked credent ia ls .
4
For m os t Am erican s tudents , then , the educ at iona l sys tem passe s the
bu ck to em ploye rs o r to the s tuden t s themse lves to p rov ide opp or tun i t i e s
to learn needed job skil ls . Employers are concerned that employees learn
the skills ne ed ed on pa rt icular jobs , ho w ev er, and there is no felt need for
their pro gr am s of on-the-job train ing to teach gen eral skil ls that can be us ed
elsew here . A nd , of cou rse, there is a lm ost nev er any formal, nat io nally
recog nized cert ification of em plo ye r-p rov ide d trainin g.
This is a very different school-work relat ionship than is found in most
othe r indu stria l societ ies , and it m ea ns that the overal l strat if icat ion proc ess
and career tra jectories are different in the U.S. than else w here . M auric e et
al.
(1986) provide an insightful analysis of the differences between the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 59/375
4
German and French systems of education as they relate to s trat if ication
processes , and Allmendinger (1989) adds fur ther unders tanding by com
pa r ing the G erm an an d Am er ican sys tems. T hese autho rs poin t out tha t the
G e r m a n e d u c a t i o n a l s y s t e m ' s h i g h l y d i f f e r e n t i a t e d s t r u c t u r e a n d
occupat ional ly- re levant credent ia ls has a s t rong capaci ty to s t ruc ture the
f low of s tudents into the labor force, while the French and American
sys tem s are very weak in this respect . Par t of the greater capa city to
s t ructure in Germany (and in Austr ia [Hal ler , Konig, Kraus , and Kurz
1985]) is d u e to the f requen t c om binatio n of school and w ork af ter leavin g
full -t ime school , thro ug h appren t iceships or o ther pro gra m s tha t sys tema t i
ca l ly combine work and s tudy.
M aur ice , e t
al.
go on to sho w h ow , in contrast to the G erm an c ase, the ve ry
gen eral creden tials of French stude nts lead to a relat ively ra n d om distr ib u
tion of stu de nt s into the labor force. Th at, in turn , lead s to a grea t dea l of job
cha ngin g in the early w ork careers , and h eavy dep end enc e on senior i ty in
th e proce ss of career mo bility. W e are left wi th a sens e of a significant
discontinuity in the careers of French young people that is not near ly as
co m m on in G erm any . M any of these sam e differences are rep orte d by
A llmen ding er in her compa r ison of G erm any and the U.S .
In many European countr ies most s tudents leave school and obta in a
full- t ime job relatively ear ly, but they con tinue to s tu dy p ar t- t im e in p ur su it
of formal, nationa lly recogn ized occupa tionally-releva nt creden tials . For
instance, the basic education-occupation l inkage in Great Britain is not as
loose as in France and the U.S. (Kerckhoff 1993). The com bi na tio n of
s p e c i a l i z e d s e c o n d a r y sc h o o l e x a m i n a t i o n s a n d n o n - u n i v e r s i t y
p o s t - s e c o n d a r y p r o g r a m s l e a d i n g to n a t i o n a l l y r e c o g n i z e d
occupationally-relevant credentials makes the careers of young Britons
almost as orderly as those in Germany and more so than in either France
or the U.S.
Most young Britons leave full- t ime school at an ear ly age, and their
educational credentials are not very predictive of their init ial labor force
placem ents . H ow ever , there is an increas ingly t ighter f it be tw een edu ca
t ional and occupational credentials during late adolescence and ear ly
adulthood in Great Britain because educational credentials gained af ter
lea vin g reg ula r school pro vi de access to spec ialized kin ds of jobs (Winf ield,
Ca m pb ell , Kerckhoff, Everett , an d Trott
1989).
A lthou gh the overa l l degree
of social m ob ility in the U.S. an d G reat Britain is hig hly s imila r, the caree r
pathways by which the mobili ty is produced are quite different.
Labor orce Structures and Stratification P roce sses
These comparative works provide clear evidence of differences in the
structures of the educational systems of industr ial societies , and they
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 60/375
4
sugges t ways in which those educat ion s t ruc tures are l inked to sor t ing
pro ce sse s in the labor force. In fact, so m e of this literatu re seem s to im pl y
that the differences in the sorting process in the labor force are simple
deriv ative s of the differences in the educa tion al system s, that any societal
differences in career patterns can be attr ibuted to educational , not labor
force, differences.
However , to in terpre t the l i te ra ture tha t way would be to over look the
great degree of in terpenetra t ion of educat ional and work exper iences in
m ost indu str ial societies . Ed uca tiona l an d labor force proce sses are not
nea r ly as wh olly sep ara te as in the U.S. Such an inter preta tion w ou ld also
imply that labor force s tructures do not vary among industr ial societies ,
5
an d the re is gr ow in g eviden ce that they do (Kalleberg
1988).
If no th in g else,
it is im po rtan t to recog nize tha t industr ial ize d societies vary in the relat ive
sizes of eco nom ic sectors (Carroll and M aye r 1986) an d occ upa tiona l gr ou ps
(Terway 1987) and in their dis tr ibutions of occupations by s tatus levels
(Ishida
1993).
To the exten t that the sh ap es of care ers differ in different p ar ts
of the labor force, these differences alone would lead to overall societal
var iat ion in career patterns.
Th ere are also societal differences in the basic dem og rap hic s of the lab or
force. Because of the low er pro po rtio ns in hig he r edu catio n and the
wid esprea d tendency to combine work and schooling , ma ny m ore you ng
pe op le are in the labor force full-t ime in m ost Eu rop ean c ou ntr ie s than in the
U.S. Th e pr op ort ion of older peo ple in the labor force also var ies s ince
retirem ent ag es an d retirem ent benefits differ across societies (O Rand and
Spi lerm an an d I shida in th is volum e) . In add i t ion, the par t ic ipa t ion of
w om en in the labor force var ies wid ely both overall an d by age an d m arital
s t a tus ,
an d de grees and kin ds of occupat ional segregat ion by gend er vary
across societies (Sorrentino 1990; Spilerm an and Ishida in this volu m e).
O the r societal dif ferences also un do ub ted ly lead to var ie d dis tr ib utio ns
of career pa tter ns. Differences a m on g socialis t , corp oratis t an d capital is t
societies are the most obvious and broadly relevant ones (Erikson and
G old tho rpe 1987; Esp ing -An ders en 1990; Treim an an d Yip 1989; Zag orski
1984). But, eve n am on g capitalis t societies , labor unio ns are m uc h stron ge r
in som e societies th an o thers (G oldth orpe 1984), an d v ar ied clar i ty of
differentiat ion among levels of occupations probably affects available
caree r pa th s (Ha ller et al. 1985).
As with the educational system, the American labor force pattern is
rat he r different from tho se of oth er ind us tria l societies. Th ere is a hi gh er
proportion of the labor force in service industr ies and occupations, fewer
youthful w orke rs , m ore w om en in the labor force , an d w eak er union s in the
U.S. than in most other industr ial ized societies . These general s tructural
features of societies labor forces are un do ub ted ly l inke d to the dy na m ics of
careers , bu t w e ha ve very l imited kn ow led ge of those relat ionship s. Tracing
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 61/375
careers thro ug h these var ied s t ruc tura l a r rangem ents requires lon gi tud inal
da ta from m ulti ple societies , an d those da ta ha ve not been w idely availab le
un ti l quite recently. W ith the increasing availabil i ty of such da ta, it sho uld
be poss ible to ma ke mo re sys temat ic and informat ive com par iso ns .
One of the diff icult ies in making progress in this way even when
ad eq ua te d ata are available is the fact that m uc h of the con cep tualiz ation of
labor force structure has been either too refined or too gross to trace career
pa tter ns thro ug h i t. Some of the m ost intr iguing co nce ptua lization s of labor
force structu re concern either bro ad ind ustria l sectors (Zucker and Rose nstein
1981) or the internal organ ization of posit ions with in f irms. Co nce pts such
as in ternal labor m arkets (A l thauser 1989), vacanc y chains (Chase 1991) and
the degree to which vacancies are op en or c losed (Sorensen and Tu m a
1981) are al l valuable contr ibutions, but they are most useful in tracing
pat te rns of m ov em ent w i thin fi rms (Rosenbaum 1984; Spi lerm an an d
L u nd e 1991), an d it is difficult to ex trap ola te from such st ud ies to na tio na l
patterns of career trajectories.
However , i t should be possible to use the logic of these approaches to
labor force analysis in developing overarching conceptualizations of ca
reers an d career l ines thr ou gh ed uca tiona l an d labor force s tru ctu res . Th ere
ha ve be en v ery few efforts to us e the con cep ts of care ers and c aree r lines in
stu die s of labor force mo bili ty. Those s tudies that hav e us ed th em ha ve
genera l ly focused on m ove m ents w i thin res tr ic ted occu pat ional ca tegor ies
(Gaertner 1980; Spilerman 1977), and efforts to expand their use to the full
ran ge of occupa tions in the labor force (Spen ner et al . 1982) ha ve de m on
strated h o w com plex the overall picture is . I t is ap pa ren t that to carry out
analyses of careers and career lines in the labor force we need a set of
catego ries that fall betw een those def ined b y jobs in interna l labor m ark ets
and industrial sectors but that the full set of occupations is overly refined.
6
At leas t two poss ible approaches have been sugges ted in previous
li terature . O ne appro ach is sugg ested by Sorensen (1977) w h o used occu
pational prestige scores rather than individual occupations as units of
an aly sis. I t is thu s possible to identify m ov es wi thi n the labor force in term s
of a s ingle hierarchical dimension whose units can be adjusted to be at
w ha tev er level of ref ined differentiat ion desire d (e.g., only deciles m ig ht b e
dist inguished and thus only ten labor force locations need be considered) .
A second approach is used by Haller et al . (1985) in their comparative
stu dy of A ustr ia , France and the U.S. They cons truct
23-category typology
of occu pat io nal /sec tora l grou ping s ( such categor ies as higher em ploy ees ,
prod uct io n; skil led w orke rs , c raft in dus tr ies ; unski l led w orke rs , bui ld ing
t r ades) an d m ap wo rker mov eme nts be tw een them over time . This second
approach is closer to that used in most s tudies of careers and career l ines,
an d I will use i t as an exam ple of a pro m ising app roa ch.
Haller et al show that the occupational categories are l inked together
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 62/375
45
du rin g the labor force careers of m en in the three societies . There are very
different f low p att ern s am on g them b etw een firs t an d later jobs. Th is is in
part a result of different distributions of the three labor forces across
occu pat ional ca tegor ies (e .g ., m an y m ore agr icul tura l w ork ers in A ustr ia
an d France , m ore ski l led wo rkers in Au str ia , m ore ma na ge rs in the U.S.) .
7
While the analyses Haller et al report are necessar i ly l imited due to the
tw o-p oin t def init ion of careers an d the age range of the sam ple s, they sho w
so m e strik ing societal differences. For insta nce , the re is essen tially no flow
between unski l led and semiski l led work pos i t ions and ski l led work pos i
t ions in A ustr ia b u t he av y f lows in France and the U.S. Also, there is mu ch
grea ter mo ve m en t across ind ustr ia l sectors in the U.S. tha n in ei ther of the
other countr ies .
The Haller et al study shows the feasibility of charting career lines
be tw een categories a nd of identifying m ajor societal dif ferences. Th at is the
bas is for sugg es t ing tha t tha t s tud y be used as a mo del . W heth er the sam e
categor ies of occup at ions are used wo uld d ep en d on the analys t s pu rpo se ,
bu t the genera l appro ach s pro m ising . I t m ay b e , for instanc e, that the social
class categories proposed by Erikson and Goldthorpe (1992) or Wright
(1985) or some other set of categories might be preferred.
C ha rtin g the f lows of w ork ers across such a set of categories d ur in g their
labor force careers is bec om ing increasing ly possible as large nation al da ta
se ts repo r t ing w ork careers becom e available. Both ind ividu al careers a nd
pro m in en t career l ines can be def ined in these terms . The logic of con cepts
such as internal labor markets and open and closed posit ions will also be
useful in disc ussin g the pa tter ns identif ied. In add it ion , i t sho uld b e
pos sible to link up these labor force careers an d career l ines with caree rs a nd
career l ines in educational inst i tutions, al though addit ional classif ication
and analys is pro blem s wil l need to be deal t w i th wh en th a t i s do ne . I wi l l
sug gest in th e nex t section so me ide as for such an an alysis for the U.S., and
in the final section I will sugg est an app roa ch to com pa rativ e an alyse s.
Educational and Labor Force
Stratification Processes in the U S
The l ite ra ture review ed ab ove sugges ts a w ay to app roach the co ncep tual
pr ob lem of l inkin g edu catio nal an d labor force s trat if ication pro cesses. Tw o
dim en sio ns of l inkag e are sug geste d. The firs t is the deg ree to w hich
educat ional ins t i tu t ions provide credent ia ls tha t have a di rec t bear ing on
ind ivid ual s occup at ional quali fica tions . The second is the deg ree to which
educa t iona l and labor fo rce ins t i tu t ions co l labora te in shap ing the
schoo l- to-work trans it ion. The U.S. evid ently has an exception ally w eak
l inkage which ever of these two dime nsion s we cons ider .
The very weakness of the linkage in the U.S. seems to lead to a set of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 63/375
6
stratif ication p rocesses that ma y be uni qu e, at least in their overall configu
ration. The exten ded p er iod of edu cation a nd the generic na tu re of our
edu cat ion al credent ia ls hav e two impo r tant imp l ica t ions . F irs t , the sys tem
essential ly rules out of any ser ious com peti t ion for goo d jobs those wh o d o
no t obtain the only basic creden tial we offer , the high scho ol dip lom a. This
is a s izable category (abou t 15% of recent cohorts) of very you ng failures
w ho are hard ly cons idered to be em ploy able in an econ om y that accepts
a 6% un em plo ym en t ra te as no rm al . A t the other end of the co nt in uu m ,
there is a ca tegory of w inn ers (college gradu ates) w ho are a lmo st gua ran
teed access to the better jobs (about 25% in recent coho rts) . In the m id dl e,
w e ha ve the major ity , an am orp ho us ca tegory of h igh school gra du ate s a nd
col lege drop-ou ts , wh ose ad ul t pos i t ions are highly p roblem at ic .
Especially for that large middle category, experience af ter school be
com es all- im porta nt in de term inin g their adult locations in the s tratif ication
system . Even af ter labor force entry, ho w ev er , the Am erican syste m seem s
to pr ov id e a less uniform set of experience s th an in mo st other societies .
Careers ap pea r to be m ore order ly an d predic table in , say, G erm any or
Jap an tha n in the U.S. (DiPrete and M cM an us and Sp ilerman a nd Ish ida in
this vo lum e). In m ost other in dus tr ial societies labor force entry po ints seem
to lead to a more l imited set of career l ines that provide order ly career
pa t tern s than in the
U.S.,
especially d u rin g the early years in the labor force.
The Am er ican sys tem seems to be especia lly op en or un s t ruc ture d in bo th
educational and labor force stratif ication processes.
Yet, recent Am erican s tud ies of both ed uca tiona l and labor force s trat i
f ica t ion processes have increas ingly emphas ized s t ruc tured sor t ing pro
cesses. In educ at ional se t t ings , w e hav e show n that gro up ing of s tu den ts
be tw een schools, be tw een c lass rooms and within c lass rooms ha s s ignif i
can t effects on their later levels of aca de m ic success (Alex and er and En tw isle
and G am oran in th is vo lume) . Al thou gh we do no t p rov ide them wi th
different kinds of credentials , our internal sor t ing processes s ignif icantly
affect bo th their acade mic ach iev em en ts (i.e. , gra de s an d test scores) a nd
their academ ic attainm ents ( i.e., high school an d college dip lom as) . I ha ve
arg ue d els ew here (Kerckhoff 1995) that these internal sor t ing m ec ha nis m s
are especially im po rtan t in the Am erican system just bec ause i t offers so few
an d such gener ic educ at ional credent ia ls . A w ard ing credent ia ls ten ds to be
an either-or matter , and the internal sor t ing mechanisms appreciably al ter
the dis tr ibutions of probabil i t ies of obtaining the credentials .
Similar ly, American s tudies of labor force processes have emphasized
such structural features as industr ial sectors , internal labor markets , va
cancy chains, and so on (Althau ser
1989;
C ha s e 1991; Ho dson and Kaufman
1984).
It is at least possible that those structural features are actually more
important in the American case than elsewhere just because our system
pas ses the bu ck from the schools to the f irms to sor t ou t the a m or ph ou s
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 64/375
7
school system p ro du cts into s trat ificat ion levels . I t m ay be that these labor
force sort ing processes are typically American processes . Or more gener
al ly, the s ignificance of such labor force sort in g processes m ay vary de pe n d
ing on the extent to which the sort ing has already been carried out by the
edu catio nal syste m . This is a t least a hy po thes is w ort h in vest ig ating .
American analys ts may face more chal lenging conceptua l and analyt ic
pro ble m s than those in other indu strial societ ies beca use the very dec entra l
ized na t ur e of our educat ion al and labor force sys tems ma kes som e of the
l inka ges be tw een th e tw o system s especial ly diff icult to identify. The re m ay
be imp orta nt bu t var ied k inds of coupl ing tha t link Ame rican educ at ional
an d labo r force ins titu tion s. If w e can identify th ese ki nd s of link ag e, w e
m ay see m or e orde r in the school-wo rk relationship than is initially a pp are nt.
For ins tance , in some American communi t ies there may be c lose l inks
betw een em ploy ers and schools tha t resemb le those found m or e gene ra l ly
in Japan . W el l -es tablished em ploy ers in som e Am erican com m un i t ies m ay
have an on-going relat ionship with the local high schools to insure a
con t inuin g su pp ly of com petent work ers . Such a re la t ionship w ou ld n ot
s tr uc tu re the f low of s tu de nts into the labor force in the sam e w ay as in
Ge rm any , bu t i t could p rov ide a mu ch m ore order ly t rans i tion from school
to w ork th an see m s general ly to occur in this cou ntry . We could expect th at
such a relat ionship would also have a feedback effect on what happens in
t h e s c h o o l . I n p a r t i c u l a r , i t w o u l d p r o b a b l y h e l p k e e p m a n y
non-col lege-bound s tudents mot iva ted to do wel l in h igh school .
The same kind of re la t ionship might exis t be tw een som e em ploy ers an d
local co m m un ity colleges or technical inst i tutes . In such cases , the college
or inst i tute curric ula cou ld ha ve a greate r capacity to s tru ctu re the f low
of s tudents into the labor force than seems general ly to be found in this
coun t ry . Em ployers migh t s eek ou t po ten t i a l emp loyees wh o hav e succes s
ful ly com ple ted par t icular k ind s of courses . Unfo r tunate ly , our k no w led ge
of vocation al po st-sec on dar y inst i tu t ions is very l imited. W e ne ed to do a
grea t dea l m ore to un der s tan d th is ub iqu i tou sbu t seemingly un s tan dard ized
elem ent of ou r educat ion al sys tem. W e hav e essent ia l ly igno red i t because
it is so difficult to m ea su re its cre den tials an d their effects on m ob ility.
But, if at least half of the non-college bo u n d hi gh school gra du ate s (Lewis ,
H e a rn , a n d Z i l b e r t 19 93 ) a n d m a n y c o l l e g e d r o p -o u t s t a k e s u c h
po st-se con da ry c our ses , the cou rses alm ost certainly affect a t least som e of
their job pro spe cts an d career patte rns . A nd if m ost of those w h o take these
cou rses are par t of w ho lly local labor poo ls , the fact that the cre den tials h av e
lit tle or no nati on al currenc y m ay no t be as im po rtan t as i t ini t ia l ly w ou ld
seem to be.
Another hypothes is worth inves t iga t ing is tha t pos t -secondary voca
t ional courses cons t i tu te a mechanism tha t feeds in terna l labor markets .
M ost of the li te ra ture on in terna l labor mark ets is concerne d w i th pa t te rn s
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 65/375
8
of m ov em en t with in firms, an d i t tend s to em ph as ize f irm-specific skil ls an d
on-the-job training (Althau ser 1989). Ho w ev er, there are also occu patio nal
inter nal labor m ark ets in wh ich gen eral ski lls and train ing are m or e clearly
involved, and post -secondary vocat ional courses may be more impor tant
there than in in-firm internal labor markets.
Even in cases in which special ized training is cal led for and nationally
rec og niz ed cr ede ntials are available, the transferability of skills in the lab or
force ma y de pe nd on the act ions of the em ploy er. For instanc e, Finlay (1983)
has shown how i t i s poss ib le for some groups of employers to l imi t the
transferabil i ty of even w idely recog nized crede ntials by l imit in g ne w hires
to those wi tho ut the formal c redent ia ls w hen ne w op enin gs occur an d n ot
pro vid ing access to the formal c redent ia ls by such unqual i f ied em ploy ees .
Courses at community colleges or technical inst i tutes could be an inte
gral part of the functioning of even in-firm internal labor markets i f
em ploy ers spo nsor the course- taking of those em ployees w ho are cho sen to
m ov e up in the f irm's hierarchy of jobs. Or, the cou rses m igh t function
indirect ly in ei ther in-firm or occu pation al internal labor mar ke ts if em plo y
ers give preference in the job qu eu e to those wh o hav e already h ad certain
kin ds of cours es. T he relevance of such cours es in the functioning of in tern al
labor m ark ets m igh t also vary dep en di ng on the kind of f irm (e.g. , large or
sm all) or the indu stry (e .g., service or m anu factu ring ) involved .
8
Our knowledge about how in terna l labor markets opera te i s not ad
equ ate to m ake very conf ident s ta teme nts abou t these m at te rs , espec ia l ly
w ith respect to trainin g cou rses that hav e only local significance. H ow ev er,
even regu lar educat iona l c redent ia ls appear to be va lu ed in d i ffe rent w ay s
by different American industr ies, as Bridges ' contr ibution to this volume
sh ow s. Similar differences un do ub ted ly exist w ith respect to voc ationa l
courses an d cred ent ia ls .
If these specu lat ions abo ut the varie d relevance of edu catio nal an d labor
force s t ruc tures for Am erican career pa t te r ns hav e any mer i t , they sugg est
that the re may be som e gain in looking for the condit io ns of that v ariat ion .
Recent research and theorizing have shown that the effects of school
s t ruc tu re var ies de pe nd ing on the k ind of school involved (Gam oran in th is
volume) or the kind of community context within which i t functions (e .g. ,
the am ou nt of c l ient po w er i t has to conten d wi th [Bidwell and Q uiro z
1991]).
Similarly, recent research and th eorizin g ab ou t labor force struc
tures have s t rongly sugge sted tha t such s t ruc tures as in terna l labor m arke ts
are found more often in some kinds of f irms and occupations than others
(Althauser 1989).
We need to assem ble mo re system atic da ta on these kin ds of variat ion so
that w e can better conc eptualize the kind s of contexts w ithin w hich sort ing
st ruc tures a re l ike ly to be impor tant and wi th in which they have varying
kind s of effects. W e pro bab ly already hav e in h an d sufficient inform ation
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 66/375
9
to warrant genera t ing some pre l iminary hypotheses , but i t wi l l take a
concerted effort to systematize that information.
n pproach to Com parative naly sis
Almost a l l compara t ive analyses d iscussed above have sugges ted tha t
the A m erica n case has un us ua l characteris t ics that sho uld affect the sh ap es
of school-work careers and, thereby, the society 's s trat if icat ion processes .
The schoo l -work l inkage appea rs
t
be we aker in the
U.S.
than in almo st any
other indus t r ia l soc ie ty . An y a t temp t to t race indiv idual career pa t te rns
across those two inst i tut ional domains in the
U.S.
leads to an im pre ssio n of
grea t variat ion . Th e discussio n in the pre vio us sect ion w as focused on
possible ways to search for some order in that variat ion.
If we assu m e a com para t ive ra ther than a who l ly Am erican pe rspect ive ,
ho w ev er, i t m ay be possible to achieve add it ion al insights into the roles of
educ at ional a nd labor force organiza t ion s in shaping careers and, the reby,
sh ap ing the overal l process of societal s tratif icat ion. O ne w ay to do this is
to cons id er the
U.S .
as an (or the) extre m e exa m ple of loos e s tru ctu ring of
careers and other industrial societ ies as contrast ing examples of kinds and
deg rees of t ighte r s t ruc tur ing .
The l i terature reviewed in this paper indicates a systematic l inkage
be tw een th e kin ds of edu catio nal and labo r force sort ing pro cesses found in
ind us tria l societies. Th ose societies in w hic h relatively refined se ts of
educ at ional c redent ia ls a re aw ard ed h ave less job chan ging d ur ing the ear ly
yea rs in the labor force an d less variat ion in w ork career mo bil i ty p att ern s .
Germany and the
U.S.
are often no ted as con trast in g cases in these re spec ts .
In add it ion , those societ ies in which there is a we ll-es tablished relat io nsh ip
betw een e duca t ional an d labor force ins t itu t ions seem to have m ore order ly
and less var ied w ork career mo bi l i ty pa t te rns . Japan an d the U.S. are often
no ted as con trast in g exam ples in this respect . The re also ap pe ar s to be a
genera l l inkage be tw een centra l control of educat ion a nd a m ore d iversi f ied
se t of na t ional ly reco gnized educa t ional c redent ia ls , a l tho ug h France seem s
to be an excep tion. Finally, the Haller et a l research s ug gests th at c learer
w ork career pa t te rn s can be found in soc ie t ies w i th h igh pro po r t ion s of the
labor force in agr icul tu re and wi th a profess ional iza t ion of m an ua l w or k
(reflected in Austria in the significance of the apprenticeship system).
These characteris t ics seem to epito m ize societ ies w ith t ight s tru ctu ring of
careers . Th ey can form the basis for a set of tentat iv e hyp oth ese s a bo ut
careers an d career l ines in ind ustria l societies . At the m os t gen eral level ,
they sugges t tha t school-work careers should be more order ly and career
l ines mo re easi ly identif ied in societ ies in wh ich there is cen tral control of the
educational system, a set of nat ionally recognized occupationally relevant
educat ional c redent ia ls , regular ized school-employer re la t ionships , and
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 67/375
5
profess ional iza t ion of m an ua l w ork .
If w e assu m e tha t the
U.S.
has a looser career s t ruc tu r ing tha n m os t o ther
indus t r ia l soc ie t ies , we would expect more ear ly job changing and more
variation in work career patterns (i .e., fewer clear career lines) in the U.S.
than almost any other industria l society. The l i terature reviewed above is
at least consis tent with tha t expectat ion. In add it ion , ho w ev er, it m ay b e
poss ible to deriv e a nu m be r of m ore refined h yp oth ese s on the basis of these
gen eral observa tions. In an at tem pt to s t im ulate greater interest in com
pa rat i ve career analysis , I w ou ld l ike to sug gest a few of these and enc ou r
age other s to d o the same. Som e of those I sug gest ha ve alrea dy received
som e supp ort in prev ious research , bu t a ll w ou ld requ ire add i t iona l inve s
t igat ion. U nd erly ing mo st of them is the basic ass um ptio n that there is
general societal recognit ion of the degree to which careers are t ightly or
loose ly s t ruc tured and tha t s tudents ' and workers ' a t t i tudes and behaviors
reflect that fact.
For severa l reasons , these hyp othes es can be v iewed as pr em atu re . To
test the m in an y systematic w ay, it w ou ld b e necessa ry to place the societ ies
dea l t wi th on the loose-t ight cont in uum and to jud ge the re la t ive im por
tanc e of the definitive criteria
I
have sugges ted , tasks tha t remain to be d on e .
Also, in al l cases , the implici t ph ras e oth er things eq ua l is im plie d, an d
tha t is bo un d to ra ise addi t iona l ques t ions , not dea l t wi th here , abo ut wh at
other th in gs need to be inc luded in tha t proviso . Ho we ver , it m ay be
preferable to be premature than to ignore the potential value of such an
effort . A s a beg inn ing , suggest the fol lowing:
(1) In societ ies in which careers are t ightly s tructured, s tudents should
hav e c learer and more rea l is t ic expec ta t ions of the i r edu ca t ion a l an d
oc cu pa tion al futures relatively early in life. Th at is , their exp ect atio ns
sho uld be closer to w ha t systematic analy ses of career pa tter ns w ou ld lead
ou ts id e observe rs to pred ict . A n analysis of the edu catio nal exp ectat io ns of
thir teen y ear old s in En glan d a nd the U.S. (Kerckhoff 1977) is consi sten t
wi th th is hyp othes is , s ince i t sho w ed grea ter rea l ism in England. H ow ev er ,
i t a lso pro vid ed evid ence of greater real ism in the
U.S.
by late ad olesce nce.
(2) If yo un g peo ple in societies w ith t ightly s tructu red careers hav e m or e
realis t ic views of the future, we might expect those who are dest ined for
relat ively low outcomes to more ful ly accept those outcomes than their
co un terp arts in societ ies w ith loosely s tru ctur ed careers . A gain us ing
En glan d a nd the U.S. as exa m ples of societ ies w ith relat ively t ightly an d
loosely s tructured careers , respectively, analysts have reported a greater
com m itm ent to ega l i ta rian a t t i tudes an d a grea ter to le rance of i r regu lar
m obi l i ty pa t te rns in the
U.S.
(Robinson and Bell 1978; Turner 1966). How
ever, s tud ies ha ve also repo rted disaffection with the educa tion al system by
som e low-p erfo rm ing adolescen ts in bo th societies (e .g. , St inch com b 1964;
Will is 1977). H ow repr esen tat ive those cases are , ho w eve r, is not c lear.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 68/375
5
(3) There should be less intergenerat ional mobil i ty in societ ies with
t ightly s tructured careers because the earl ier individuals enter career l ines ,
th e gre ate r the effects of social origin on caree r ent ry po rta ls is likely to be .
It
s
f requently rep ort ed that the effect of social origin on edu catio nal success
declines as s tudents move through the levels of education (Shavit and
Blossfeld 1993; H ou t in this vo lum e), so the t ighter the l inkag e bet w een
early and later posi t ions in the educational system (and then in the labor
force),
the less social mobil i ty there should be.
(4) Internal labor markets should play a more sal ient role in the careers
of w or ke rs in societ ies w ith loosely s tru ctu red careers . Researc h tha t
repo r ts m ore wi th in-f i rm adv ance m ent in France than in e i ther G erm any or
E ng lan d (M ar sd en 1986; M aur ice et al 1986) is at least con sisten t wit h this
hyp oth esis (see also Kalleberg and L incoln
1988;
Loveridge
1983;
Stark 1986).
C o n c l u s i o n
Th e m os t genera l poin t in th is chap ter is tha t we ne ed to know m ore than
we do about the ways in which educat iona l and labor force ins t i tu t ions
opera te toge ther to shape individual careers and, in so doing, se rve to
s t ra t ify the m em be rs of the soc ie ty . On e sugge s t ion— derivab le f rom
M aurice , e t a l. , A l lme ndin ger , and o thers—is tha t edu ca t iona l ins t i tu t ions
are relat ively m or e im po rta nt in som e societies an d labor force i nst i tu t ion s
are relat ively more important in others . I t seems l ikely, however, that the
s i tua t ion is even m or e complex than tha t .
Especial ly in the United States , there seems to be the need for two
add i t iona l k in ds of ana lys is . Fi rs t, w e hav e not ye t ade qua te ly ind exed the
role of educa t iona l ins t i tu t ions because we have a lmos t whol ly ignored a
major par t of the pos t -secondary exper iences of our popula t ion—courses
taken in com m un i ty col leges and technica l ins t i tu tes . W e hav e igno red
these courses in la rge pa r t for m ethodo logica l reasons— we d on t kn ow
ho w sys temat ica l ly to take them in to account . But they un do ub ted ly m ak e
a great de al of difference in the sort ing processes in this society an d w e n ee d
to specify their role.
Secon d, w e hav e yet to do an ad eq ua te job of accou nting for the k ind s of
l inkag e be tw een edu catio nal and labor force inst i tut io ns. W e ha ve effec
t ively treated those l inkages as if they do not exis t because they are not
sufficiently reg ula rize d to just ify us ing the m in a cons is tent ma nn er in ou r
ana lyse s. It see m s likely that the re is a var iety of type s of link ag e and tha t
the several types wil l be found in different kinds of contexts . We need to
attempt to identify the kinds of contexts in which the several types of
l inkages m os t f requent ly app ear .
We also need to think through more carefully than we have thus far the
theoret ical implicat ions of our usual views of the ways s tructural effects
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 69/375
52
com e abou t in edu catio nal and labo r force contex ts . There are tw o basic
differences betw een o ur ideas abo ut the natu re of s tructu ral effects in those
two sett ings. Firs t , most s tructural arrangements we study in schools are
seen as purposely established to affect the achievement levels of the
s tud en ts . W e m ay believe tha t they do not a lways accomplish the ir
pu rpo se , bu t they are seen as a t tem pts to influence s tud ent ou tcom es . In
con trast , the labor force s tructure s we mo st of ten s tud y are seen as ha vi ng
be en de sign ed to increase the effectiveness of the firms. To ap pre ciate the
possible s ignif icance of that difference, i t may help to imagine how our
educational system might be affected (and how our theories of s tructural
effects w o u ld ne ed to be altere d) if all of
school s opera t ing reso urces we re
determined by the collective performance of i ts s tudents .
Second , exp lana tions of the effects of s tructu ral arra ng em en ts in the tw o
settin gs focus on very different k ind s of pro cess es. In scho ols, th e st ru ctu res
are seen as directly affecting s tu de nt s ach ievem ents thro ug h cu rr icula,
teachin g s tyles , social def init ions, and s tu de nt m otives. In con trast , in the
labo r force, differences in w ork er ou tco m es are seen as the indire ct resu lt of
f irm or ind us try organ ization al constraints an d cost-benefit con sidera tion s.
In ad dit io n to these being qu ite different c onc eptu alizatio ns, bo th of th em
may be over ly nar row.
The most basic point about the American case is that the decentralized
control of educat ion an d the re la tive au ton om y of em ployin g organ iza t ions
h a v e p r o d u c e d a h i g h l y v a r i e d s et of e d u c a t i o n a l o u t c o m e s a n d
educa t ion-occup at ion re la t ionships , and we need to dev ote some t ime and
ene rgy to searching for some order in w hat a pp ear to be highly disp ara te
stratif ication pro cesses. In do ing so, w e will prof it f rom usin g a com pa ra
t ive persp ective . I t wil l pro vid e clues as to possib le al ternative w ay s of
stru ctu r ing s trat if ication processes, an d i t will m ak e us mor e sensit ive to the
unique features of the American case.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 70/375
53
Notes
1.1 am ind ebte d to David G rusk y for helping m e recognize this w ay of orga nizi ng
the l i terature on strat i ficat ion processes .
2.
On e of the resul ts of the lack of different iated crede nt ials is a st ro ng em ph asi s
in the U.S. on w h er e stud en ts go to school or col lege.
3.
It is very difficult to obtain reliable inform ation ab out the cour ses tak en or th e
credent ia l s ob ta ined in pos t -secondary prog ram s in com m uni ty co l leges and tech
nical inst i tutes. For instance , in the 442 pa ge pub licat ion of the Na tiona l Cente r for
Edu cat ion Stat istics, The Co ndit ion of Edu cat ion
1994,
there is no me ntion of these
k ind s of courses . T he on ly index en t r ies un der technica l and voca t iona l a re
concerned with high school and regular col lege courses.
4. The k inds of c redent ia l s Al thauser and Appel d i scuss in th i s vo lume are an
except ion, in that they are nat ional ly stan da rdiz ed a nd recognize d. But they are
crede nt ials for semi-professional occu pat ions , and the y are not typical of cr eden t ials
earned th rough post -secondary voca t iona l courses .
5. That implicit assu m ptio n of uniformity of labor force st ruc ture s is consistent
w i th mu ch of the comp ara t ive l i t e ra ture on occupat ional m obi li ty . There
is
ev idence
of a co m m on hierarchical or der of occupat ions (Treiman 1977) as we ll as similari ty
in the amount of intergenerat ional mobil i ty in industrial societ ies (Ganzeboom,
Treiman, and Ultee 1991). However, inter-societal uniformity in mobil i ty pat terns
is increasingly being quest ion ed (Gan zeboom , Luijkz, and Treim an
1989;
European
Sociological Re view , Vol.
N o. 3 [December
1992];
H ou t in th i s vo lum e) , and grea ter
at tent ion is bein g directed to w ard un de rst an din g intersocietal differences.
6 . In an at tempt to l ink educat ional locat ions with labor force locat ions, I have
used
a
four-pa rt classification of indu strie s (the cross-classification of cor e-p eriph eral
with service-product ion (Kerckhoff
1993),
but i t prove d to be overly crud e. A m ore
ref ined scheme i s needed .
7. The y also ack no w ledg e a classificat ion p roble m . Since so m an y A ustrian m en
enter the labor force as apprentices, they used the men's first job after the end of the
app ren t ices hip as their fi rst job. This essentially classifies apprenticeships as a form
of edu cat io n. Tha t special feature of the Austrian data an d their wa y of hand ling i t
clearly reflect the differences already discussed between stratification regimes in
which there is an interpenetrat ion of educat ional and work experiences (such as
Aust r ia and G erm any) and those in wh ich the tw o are a lmost comple te ly s epara te
(such as the U.S.) Th e cou ntries ' m eth od s of reco rding wo rk ers ' posi t io ns in the
labor force also reflect such differences. In U.S. cen sus data, appr en tice s are
classified as op era t ives . If Brit ish data w ere includ ed, com par at ive analysis
would be even more difficult because the British classify apprentices in the occupa
t ions for which they are t raining.
8. There could be a direct l ink between kinds of f i rms and the courses workers
take in tha t som e em ploye rs may o penly encou rage and even f inancia lly sup por t
cour se-tak ing. Or the re could be an indirect l inkage. For instance, i t m ay be that
when there a re employers in the communi ty who reward course- tak ing , there i s
m ore publ ic und ers tan din g of the po ten t ia l va lue of such courses , and peop le take
them on their own ini t ia t ive.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 71/375
5
e f e r enc e s
Allm end inge r, Jut ta . 1989. Care er Mobil i ty Dy nam ics - A Co m pa rat iv e An alysis
of the Uni ted Sta tes , N orw ay, and W est Ge rma ny.
Berlin:
Ma x-Planck Ins t i tut fur
Bi ldungsforschung .
Al thauser , Rober t
P .
1989. Internal Labor M arke ts. A nnu al Review of Sociology
15:143-61.
Bidwell , Ch arles E., and Pamela Qu iroz. 1991. Org anizat iona l Control in the Hig h
School W orkplace : A Theore t ica l A rgum ent . Journa l o f Research on Adoles
cence
1:211 29.
Ca rroll , G lenn R., an d Karl Ulrich M ayer. 1986. Job-Shift Patte rns in the Federa l
Rep ublic of G erm any : The Effects of Social Class, Industrial Sector an d O rgan i
zat io nal Size. Am erican Sociological Review
51: 323-41.
Ch ase, Ivan D. 1991. Vacancy Ch ains . A nn ual Review of Sociology 17:133-54.
Erikson, Robert , and John H. G oldth orpe . 1987. Co m m ona li ty an d Va riat ion in
Social Fluid ity in Ind ust rial N atio ns. Pa rt II: The Mo del of C or e Social Fluidity
A pp lied . Eu rop ean Sociological Review 3:145-66.
. 1992. Constant
Flux:
Co m par at ive Analysis of Social Mob il ity in Indus trial
Nat ions . Oxford : Clarendon Press .
Es ping- A nde rsen , Gosta. The Thre e W orlds of W elfare Cap ital ism. Princ eton, NJ:
Princeton Universi ty Press.
Finlay, W. 1983. O ne Oc cupa t ion, Tw o Labor M arkets: The Ca se of Long-Shore
Cr an e O per ator s. Am erican Sociological Review 48:306-15 .
G aertne r, Karen N . 1980. The Structure of O rganiz at ional Ca reer s, Sociology of
Educat ion 53: 7-20.
G an ze bo om , H ar r y B. G. , Ru dd Lu i jkx , and D ona ld T . Tr e im an . 1989 .
Interge nerat io nal Class Mobil i ty in Co m par at ive Persp ect ive. Research in
Social Stratification and Mobility 8: 3-84.
Ganzeb oom , H ar ry B .G. ,Dona ldJ. T re iman and W oo tC .U l t ee . 1991 . Com para t ive
Intergen erat ional Strat if icat ion Research: Three Ge nerat io ns an d Be yond.
A nn ua l R eview of Sociology 17: 277-302.
G old tho rpe , John H . 1984. The End of Co nver genc e: Co rpora t ist an d Dualist
Ten dencie s in M ode rn Western Societies. Pp . 315-43 in J. H. G old th orp e (ed .)
O rd er and Conflict in Co ntem por ary Capital ism. Oxford: Oxford U nivers i ty
Press.
Haller, Max, W olfgang Konig, Peter Kra us, and Karin Ku rz. 1985. Pa t tern s of
Car eer Mobil i ty an d Structural Posi t ions in A dva nce d Capital is t Societies: A
Co m pariso n of M en in Au st r ia , France and the Uni ted Sta tes . Am erican
Sociological R eview 50: 579-603.
Ha ll inan, Mau reen T. 1992. The Or ganiz at ion of Stu den ts for Instruct ion in the
M iddle School . Sociology of Edu cat ion 65:114-2 7.
H od so n, R., and R. L. Kaufm an. 1984. Econo mic Du alism: A Cri t ical Review .
American Sociological Review 47:272-39.
Ishida, Hiro shi . 1993. Social Mob il ity in Co nte m por ary Japan . Stanford, CA :
Stanford Universi ty Press.
Kal leberg , Ar ne L. 1988. Co mp ara t ive Perspec t ives on W ork St ruc tures and
Inequ al i ty . A nn ua l Review of Sociology 14: 203-25.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 72/375
Kalleberg, A m e L., and Jam es R. Lincoln. 1988. The Stru cture of Earn ings
Inequa l i ty in the U.S. an d Jap an. A m erica n Jou rnal of Sociology 94 (Sup ple
ment ) :
121-53.
Kerckhoff
Alan C. 1977. The Real ism of Educ at ional A m bit ions in En gland and
the U nited State s. Am erican Sociological Review 42:
563-71.
. 1993. Diverg ing Pa thw ays : Social Stru cture and Care er Deflect ions. N ew
York: Cambridge .
. 1995. Reform ing Educat ion : A Cri tical Ov erlooked Co m pon ent . For th
c o m i n g in M a u r e e n T . H a ll in a n (ed.), Restru cturing Schools: Pro m ising P ract ices
and Policies. New York: Plenum.
Lareau, A nne t te . 1989. Ho m e A dva ntag e: Social Class and Parental Interve nt ion
in Elem entary Educ at ion. Lon don: Falm er Press.
Lew is, Darrel l R., Jam es C. Hearn , and Eric E. Zilbert. 1993. Efficiency and Eq uity
Effects of Voc at ional ly Focused Postse cond ary Ed ucat io n. Sociology of Educ a
t ion 66 :188-205.
Lo verid ge, R. 1983. Sources of Diversi ty in Internal Lab our M arke ts. Sociology
17:44-62.
M ars den , D. 1986. The End of Econom ic
Man?:
Cu stom and Co mp et i t ion in Labour
M arkets. N ew York: St . M art in 's Press.
M au rice , M arc, Fra nco is Sellier, an d Jean- Jaque s Silvestre. 1986. Th e Social
Fou ndat io ns of Indust r ia l Power: A Com parison of France and Ge rm any. Ca m
bridge , Mass . : MIT P ress .
M uller, W alter, an d W olfga ng Karle. 1992. Social Selection in Ed uca tion al Sys tem s
in Eu rop e. Eu rop ea n Sociological Review 8:233-54.
Raffe, Dav id, and N . Tom es. 1987. The Organ isat ion a nd Co nten t of Stud ies at the
Post -C om pulsory Level: Co unt ry Study: Scot land . OE CD Educat ional M ono
grap h . Par i s: OE CD .
Ro binso n, Rob ert V., an d W end ell Bell. 1978. Eq uality , Success, an d Social Justice
in En gland an d the Un ited States. A me rican Sociological Review 43:125-43.
Ro sen bau m , Jam es E. 1984. Career Mobil i ty in a Co rpo rate Hierarc hy. N ew Y ork:
Academic Press .
Ro sen bau m , Jam es E., Tak ehiko Kariya, Rick Set tersten, and Tony M aier. 1990.
Market and Network Theories of the Transi t ion from High School to Work:
Their Applicat ion to Indu strial ized Societ ies. A nnu al Review of S ociology 16:
263-99.
Shavit, Yossi, and Hans-Peter Blossfeld (eds.). 1993. Persistent Inequality: Chang
ing Educat ional At ta in m ent in Thi r teen Cou nt r ies . Boulder , CO : W esrv iew
Press .
Soren sen, A age B. 1977. The Struc ture of Inequ al i ty and the Process of A ttain
m en t. A m eric an Sociological Review 42:965-78.
Sorens en, A age B. an d N an cy B. Tum a. 1981. Labor M ark et Stru ctures and Job
M obility . Res earch in Social Stratification an d M obility
1:67-94.
Sorren t ino , Con stance . 1990. The Ch ang ing Fami ly in In terna t iona l Perspec t ive .
Monthly Labor Review 41-58.
Spenner, Kenneth I . , Luther
B.
O tto and V au gh n R. A. Cal l . 1982. C areer L ines and
Ca reers. Lexington , Mass.: Lexington Books.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 73/375
6
Spiler m an, Sey mo ur. 1977. Care ers, Labor M arket Structure , and Socioeconom ic
Ac hievem ent . Am erican Journa l of Sociology 83:551-93.
Spi le rm an, Seymou r, an dT or m on d Lun de . 1991. Fea tures of Educat ional At ta in
m en t and Job Pro m otion Prosp ects . Am erican Journal of Sociology 97:689-720.
Sp ring , Joel. 1976. Th e So rting M achine. N ew York: Da vid McK ay.
St inchco m be, A rth ur L. 1964. Rebel l ion in a H igh School . Ch icago: Q ua dr an gl e
Books.
Stark, Dav id. 1986. Rethin king Internal Labor M arkets: N ew Insights from a
Co m pa rat i ve Persp ect ive. Am erican Sociological Review 51: 492-504.
Terw ey, M. 1987. Class Posi t ion and Incom e Inequal i ty: Co m par ing Resul ts for the
Fede ral Repub lic of G erm any with Cu rren t Research in the US A. Intern at ional
Journal of Sociology 17:119-71.
Tre iman, Donald
T.
1977. Occ upat ional Prest ige in Co m par at ive Persp ect ive. N ew
York: Academic Press.
Treim an, Do nald T., and Kam-Bor Yip. 1989. Edu cat ional and O ccup at iona l
At ta inment in 2 Co untr ies. Pp. 373-94 in Me lvin L. Kohn (ed.) , Cro ss-N ational
Research in Sociology. N ew bu ry Park , CA : Sage.
Tu rner, Ralph H. 1966. Acc eptance of Irregular Mobil i ty in Bri tain an d the United
Sta tes. Soc iom etry 29:334-52.
Us eem , El izabeth L. 1992. M iddle Schools and M ath G rou ps: Pa ren ts ' Invo lvem ent
in Ch ildr en 's Placem ent . Sociology of Edu cat ion
65:
263-79.
W ill is, Paul . 1977. L earning to Labor: H ow W orking C lass Kids Get W ork ing C lass
Jobs.
N ew York: Colu mb ia Universi ty Press.
W infield, Ide e, Richard
T.
Campbel l , Alan C.
Kerckhoff
Diane D . Everet t , and Jerry
M. Tro tt . 1989. Ca reer Proc esses in Gr eat Britain and the U nited States. Social
Forces 68: 284-308.
W rig ht, Eric O. 1985. Class es. Lo ndo n: Ve rso.
Z a g o r s k i ,
Krysztof.
1984 . C om pa ris on s of Social M obi l i ty in Differen t
Socio-Economic Sys tem s. Pp . 13-41 in Manfred Niessen, Jules Pesch ar, and
Ch anta l Kouri lsky (eds.) , Interna t ional Com par at ive Research: Social Struc tures
and Public Inst i tut ions in Eastern and Western Europe. Oxford: Pergamon.
Zuc ker, Lyn ne G. and Caro lyn Rosenstein. 1981. Tax onom ies of Inst i tut ional
St ru c ture : Dual Econom y Rec onsidere d . Am erican Socio log ical Review
46:869-84.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 74/375
PART TWO
Educational ontexts
and Processes
The pape rs in Par t II he lp us und ers tan d h ow e duca t ional se t t ings an d
processes genera te a cumula t ive d ispers ion of s tudent performances , a
major con tribu tor to social s trat ificat ion. The firs t two pa pe rs deal w ith the
formal org aniz ation of s tu de nt s within schools . In the f irst of these, A d am
G am or an rev iew s the s tate of our kno w led ge abo ut the effects of org an izin g
s tud en ts in such ca tegor ies as abi li ty gro up s and t racks . The genera l pa t te rn
he rep orts is for those in hi gh gr ou ps to gain academ ically from tha t
placem ent bu t for those in low gro up s to fall back academ ical ly . Such
placements also affect the at t i tudes and expectat ions of s tudents and their
s ignificant other s . Since pla cem ents tend to pers is t ove r t im e, they ap pa r
ent ly ha ve cum ula t iv e effects on outco m es . G am ora n reviews severa l
theoret ical approachs to interpret ing these effects , and he recognizes the
probable importance of both symbols ( the social meaning of locations in
school s tructures) and inst i tut ional practices ( teaching s tyles , curricular
offerings).
Kar l Alexander and Doris Entwis le fo l low Gamoran 's genera l review
with a report on the effects grouping by abil i ty in f irs t grade have on
outco m es in m idd le school . Their de ta i led d a ta come from a longi tu dina l
s tu dy of s tu de nt s in Balt imore inne r ci ty schools . They inc lud e in their
defini t ion of org aniz ation al differentiat ion no t only abili ty gr ou ps and
tracks bu t also special edu catio n classes and r etentio n in gra de . All of these
make c lasses more academical ly homogeneous , and a l l of them have
identif iable effects on outc om es in s ixth gra de. This f ine-grained s tu dy
pro vid es clear eviden ce tha t som e s tud ents exper ience a do w nw ar d spi ra l ,
not only in their academic performance, but also in their a t t i tudes toward
school and their academic self- images.
The third p ap er in Part II, by Charle s Bidwell , Step hen P lanck, a nd
C ha nd ra Mul ler , takes w hat O 'Ra nd refers to in Ch apte r On e as a re la t ional
app roac h to un der s tan d in g s tude n t s ' ca reers . They show how the f r iend
ship networks of high school s tudents mediate the effects of schools ,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 75/375
58
famil ies and neigh borh oo ds on s tu de nts aspira t ions and a t t i tud es as wel l
as their par t ic ipatio n in school an d wo rk activit ies . The ne tw or ks a re
viewed as sources of both informat ion and normat ive def ini t ions and
sanctions, and they help us to interpret the effects of school, family and
ne ighborhood .
The last two papers in Par t I I are concerned with kinds of educational
reforms a nd their effects on equity and s t ud en t ach ievem ent. J. D ou glas
Wil lms examines what happened when Scot t ish parents had a choice of
schoo ls for their childre n to at ten d. H e asks wh eth er the practice inc reased
the deseg rega tion of social s tatu s gro up s, as som e sup po rter s of the practice
hav e predic ted? His da ta come f rom Scot land, bu t the observed p at tern s are
prob ably not uniq ue to tha t coun try . As man y have predic ted , requ es ts for
par t icular schools in Scotland init ial ly came heavily f rom middle class
parents , and the schools most f requently requested were largely in high
stat us areas. Th e overall pa ttern across a ten-year per iod of school choice
was an increase in the segregation of middle class s tudents , most clear ly
seen in large urba n areas. School choice clearly did not redu ce social s tatus
segregat ion.
In the f inal pap er in Par t II, M au reen H all inan no tes that prop osa ls for
school reform are f requently generated by concern with broader social
issu es— w ith value conf licts, economic problem s and so on. She arg ue s tha t
there sho uld be one pr im ary bas is for evalua t ing prop ose d reforms, the i r
exp ected effect on s t ud en t learning , beca use tha t is the basis of e du cati on al
achieve me nt and la ter occupat ional p lacement . Af ter analyz ing the three
pr im ary co ntr ib utors to s tu de nt learning (abil ity, effor t, and op po rtun it ies
to learn) , she discusses several reform proposals (such as magnet schools ,
do ing aw ay w ith cur r iculum tracks , and school- to-work prog ram s) in te rm s
of bo th the forces that br ou gh t the m forw ard an d their probab le effects on
s tud ent learning and achievem ent . She points out tha t even reforms with
no n-l ear nin g goals can ha ve posit ive (or neg ative) effects o n s tu de nt learn
ing.
These f ive papers deal wi th severa l mechanisms and processes tha t
either directly or indirectly affect the dis tr ibu tion of acad em ic ach ievem ent.
In the pa pers by Gam oran an d by Alexand er and Entwis le the me chan isms
are the intern al organ ization s of schools and the categorical differentiat ions
am on g s tud en ts . Bidwel l e t a l . show h ow netw orks of peer re la t ions serve
to mediate the inf luences of schools , families and communities on s tudent
a t t i tudes and behav iors . An d Wil lms and H al l inan discuss the wa ys in
which policy decis ions can and do alter academic sett ings and s tudent
perform ances. Tog ether , these pa per s sugg est the array of sources of
inf luence o n ed uca tion al s trat if ication.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 76/375
Educational Stratification
and Individual Careers
Adam
amoran
University of Wisconsin Madison
M y aim in this ch ap ter is to describe an d assess research o n the s trat if ied
s t ruc tu re of edu cat iona l sys tems . I am pr im ari ly concerned wi th s tu dies of
the impact of var ied exper iences wi th in schools on cogni t ive and s ta tus
ou tcom es. I wil l no t ad dr ess the full spe ctru m of w or k in the s tatus
a t ta inm ent t radi t ion— that i s, research on the de te rm inan ts of edu cat ional
and occupational success—but wil l concentrate on the subset of this l i tera
tur e w hic h exam ines s trat ification within levels of sch oo ling, and the effects
of these h ierarchies on the prog ress of indiv iduals th rou gh the edu cat ion al
sys tem. Clear ly , research on s ta tus a t ta inm ent has pro vid ed a n essent ia l
s t im ulu s to s tudi es of s tratif icat ion in school syste m s, an d
will take no te of
ho w theore t ica l de velo pm ents an d empir ica l d i rec t ions in the s ta tus a t ta in
ment tradit ion have influenced work on s trat if icat ion within school sys
tem s. In ad dit io n, I wil l sho w tha t research in this area has thre e i m po rta nt
characteris t ics that ha ve earne d i t sub stantia l a t tentio n w ithin the sociology
of ed uca tion : I t is cum ula tive , in that s tudie s tend to bu ild o n one ano ther ,
leading to subs tant ia l gro w th in kn ow ledg e; i t i s com para t ive , w i th s tudies
from severa l o ther countr ies complement ing work done in the Uni ted
States; an d i t is policy-re levant , s t im ulate d by and c on tribu ting to d eba tes
about how school ing should be organized.
Effects of Stratification in Schools:
Socializatio n A llocation and Legitimation
Sociologis ts have long recognized tha t persons who accumula te more
yea rs of schoo ling at tain high er s tatus as ad ults e .g., Soro kin 1927). At each
level of school ing, however , cons iderable var ia t ion in s ta tus a t ta inment
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 77/375
60
rema ins . At one t ime the convent ional w isdom held tha t the ma in reason
for th is d ispar i ty w as inequal i ty amo ng schools : s tud ents w ho com pleted
their ye ars of s tu dy at a go od school w ere better off tha n thos e w h o ha d
go ne to a ba d school, other thing s bein g equ al. By the ear ly 1970s,
ho w ev er , i t w as clear that this exp lanatio n w as largely inad eq ua te: Differ
ences among schools had l i t t le impact on var ia t ion among s tudents , wi th
other con di t ionsh eld co ns tant (Coleman e t al. 1966;Mosteller an d M oy nih an
1972;
Jen ck s et al. 1972; A ve rch et al. 1972).
One important reaction to this f inding was the decis ion to consider
var ia t ion in s tu de nts ' exper iences insi e schools. Because s tuden t outc om es
var ied m uc h m ore wi thin schools than betw een the m, it m ad e sense to look
at the connection between within-school differences in schooling and
student outcomes. Seminal contr ibutions to this effor t include s tudies by
H ey ns (1974) an d Alexa nde r an d McDill (1976) , w h o sho w ed th at cu rr icu
lum t racking in high schools w as associa ted w ith var ia t ion in e duca t ional
resul ts . S tud ents enrolled in a college prep ara tory cur r iculu m w ere m ore
likely to pla n on atten din g college, to earn hig h grad es, and to score hi gh on
tests.
Th ese w ere not the f irst surve y s tudie s of curr i culu m tracking, b u t th ey
w ere th e first to esti m ate effects of track ing hol di ng co nst an t differences in
st ud en ts ' abil it ies pr io r to tracking. They ref lected and co ntr ibu ted to bo th
the school-effects an d the s tatus at tainm ent trad it ions. Both s tud ies rel ied
on a conceptual model of the school as a socializing agency, in which
college- track me m be rsh ip pro vid ed a context for social interactions that led
to hi gh er expe ctatio ns an d hig he r test scores. By the en d of th e 1970s,
however , two impor tant theore t ica l a r t ic les had added new views of the
school in the at tempt to understand the role of schooling in s tatus at tain
m en t: Kerckhoff (1976) em ph asiz ed the
allocative
role of the scho ol, a nd
Meyer (1977) focused on the school as a legititnizer
K e r c k h o f f (1 97 6) r e i n t e r p r e t e d t h e a s s o c i a t i o n b e t w e e n
social -psychological processes and a t ta inm ent . He argue d tha t chan ges in
s ta tus am bit ions associa ted with school ing reflected s t ud en ts ' respo nses to
status classif ications and resource al locations rather than interactions that
s tu den ts exper ienced in school:
In the school setting, for example, teachers make decisions when assigning
grades, dividing a class into reading groups or other functional units, or
singling out individual students for special attention....These decisions not
only provide the individuals involved with information about themselves
and their probable future, bu t they also create socially significant classifica
t ons on the basis of which others w ll respond to them differentially. In short,
the decisions segmentalize the population of studen ts into categories whose
attainm ent p robabilities are different. The same kinds of decisions are mad e
by counsellors...(p. 374-375).
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 78/375
61
This conceptual framework redirected the at tention of sociologis ts to
w ar ds a l loca t ive processes occurr ing in schools, and the i r im pact on out
comes for s tudents .
A year la ter , M eyer 1977) arg ue d that bo th socializat ion an d al location
co ul db e sub sum ed un de r a theory of edu cat ion as a legi t imiz ing ins t i tu t ion .
For M eyer, schoo ling affects s tatu s at tain m ent by creating rules an d un de r
s tandings in which persons in some ca tegor ies a re e levated above others .
A n edu cat iona l sys tem not only a l loca tes pers ons to pos i t ions ; i t a lso p lays
a key role in providing the authori ty and legit imacy that defines the
hierarchy to which those persons are al located.
For the mo st part , sub seq ue nt research h as no t led to the reject ion of on e
view social izat ion, al location, or legit imation) over ano ther . Instea d,
scholars have examined al l three aspects of the role of schooling in s tatus
at tainment. Studies of s trat if icat ion in school systems—including work on
spec ia l i zed sch oo ls , cu r r i c u lu m t racks , ab i l i ty - g ro upe d c la s ses , an d
within-c lass abil i ty grou ps— hav e d raw n on a l l three of these perspect ives ,
al though usually without explici t ly contrast ing them in the context of a
s ingle s tud y. Th us , theoret ica l develop m ents in s ta tus a t ta inm ent research
in the 1970s set the agenda for work on stratification in school systems
occurr in g over the next two d ecades .
Stratification and Socialization in Schools
Studies tha t emphas ize soc ia l iza t ion as a key mechanism re ly on e le
ments of the al location and legit imation perspectives to set the s tage for
di f ferentia l soc ia l izat ion. A pr im e exam ple com es f rom wri te rs wh o
present effects of t racking on expectat ions, a t t i tudes , or achievement, and
describe these as the outcome of varied social izat ion across tracks: a long
side the at ten tion to social izat ion for varied ou tcom es, these au tho rs recog
nize the im por tance of a l loca t ion processes in de ter m ining h ow s tud en ts a re
div ide d into the different t racks. O nly a few writer s ack no w led ge this
explici tly e .g., A lexa nd er, Co ok, an d McD ill
1978;
Gam oran an d M are 1989 ;
Kerckhoff 1990), b u t the view is imp licit in all suc h stu di es. M or eo ve r,
researc h in this area accepts the im po rtan ce of extern al no rm s in ass ignin g
legit imacy to the s trat if ied s tructure of school systems, thus set t ing the
br oa d context for gr ou pi ng and tracking G am ora n and Berend s 1987). In
this sense, then, s tudies of s tatus at tainment as social izat ion invariably
d ra w o n al location and leg it imation theories . St il l, research in the social iza
t ion tradit ion is dis t inc t ive in i ts ass um pti on th at var ied outc om es a re
pr od uc ed not by the me re a l loca t ion to var ied po s i t ions , bu t by soc ia liz ing
experiences that occur within the different schools , t racks, and groups.
Group ing tracking and achievement. M any wr i t e r s hav e sho w n tha t
s tud ent s ass igned to h ighe r-s ta tus abil ity gro up s an d curr icu lum t racks
ga in m ore know ledge an d sk il ls , a s m easured by s tanda rd ized ach ievement
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 79/375
62
tests, com pared to studen ts in lower tracks Oakes, Gam oran, and Page
1992). Several recent stud ies explore the socializing experiences tha t yield
such achievement differences; these studies emphasize variation in class
room instruction across high, midd le, and low groups an d tracks. In
elem entary schools, Barr and D reeben 1983), Rowan and Miracle 1983),
and G am oran 1986) observed that the faster pace of instruction for
high-ability reading groups contributed to achievement advantages over
the course of
a year.
At
the
secondary level, stude nts
in
college-preparatory
programs take more academic courses, and especially more advanced
academic courses in ma th and science, and these experiences foster achieve
m en t gains Gam oran 1987; Sebring 1987). Teacher surveys suggest that
honors math and science classes in middle schools add to achievement in
pa rt by introducing stud ents to more problem -solving activities Hoffer
and Gamoran 1993), and an observational study indicated that greater
attention to serious academic content gave honors English stude nts a n ed ge
over their regular- and remedial-class cou nterparts Gam oran et al., in
press).
In an imp ortant rev iew, Slavin 1990) issued a challenge to sociological
studies show ing increasing achievement gaps am ong stud ents assigned to
high- and low-status classes. Slavin examined a large body of w ork on the
effects of ability-grouped classes in secondary schools, and concluded tha t
inconsistencies in estimated effects of grouping, which centered around
zero, mean t that grou ping has no effects on achievement. No t only were the
effects zero on average, Slavin argued, but high-group students did not
gain, and low-group stu den ts did not
lose,
as a consequence of their v aried
pos itions in the hierarchy . Slavin offered two explanations for why stud ies
of high and low tracks tend to show increasing
inequality:
First he allowed
that broa d curriculum tracking— that is, assignment to separate prog ram s
which dictate diverse arrays of courses—might have cum ulative effects on
achievement not appare nt in nar row er stud ies of ability grouping . Second,
he argued that observed effects of ability grouping in non-experimental
stud ies reflected selection bias—that is , effects were caused by unexam ined
differences among students in varied classes, rather than by differences
among the classes themselves.
have offered a different in terp reta tion for inconsistencies in research on
ability gro up ing Gam oran 1993). My view is that small-scale natura listic
and experim ental stud ies are likely to differ in the way g rouping is imple
mented; in particular, they may differ in how instruction is allocated to
varied
classes.
In some experiments, teachers may have provid ed the same
instruction to all classes; this wo uld likely result in no a ppare nt effects of
grouping. In other studies, teachers may have worked hard to remediate
those in low-ability classes, bu t still other cases may exist in which bette r
instructional resources were directed tow ards high-ability classes. My
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 80/375
63
interpretation is actually consistent with Slavin's (1990, p. 491) general
conclusion that un less teaching m ethods are systematically chan ged ,
school organization has
little
impact on student achievement. On average,
in the real world, research indicates that instructional content and m etho ds
do vary systematically, favoring those in high-status classes and working
against those at the bottom of the hierarchy (Gamoran and Berends 1987;
M urp hy and Hallinger 1989; Oak es, Gamo ran, and Page 1992). For this
reason, I m aintain, naturalistic studies of representative samples are m ore
likely than small-scale experiments to show increased inequa lity associated
with ability group ing.
My conclusion is suppo rted not only by observational evidence, but also
by the only two studies of ability group ing using national survey d ata that
permit comparison of students in high- and low-ability classes to compa
rable stude nts in mixed-ability classes. In a study of British secondary
schools, Kerckhoff 1986) found that students in high-ability classes gained,
and those in low-ability and rem edial classes lost gro und, relative to those
in un gro up ed classes. In the United States, Hoffer 1992) observed a similar
pattern for junior high math and science achievem ent. Moreover, Hoffer
1992)
replicated
his
math and science findings using
a
statistical mod el that
adjusted for possible selection bias in estimates of track effects. Using the
same technique, Gamoran and Mare
1989)
reported that earlier studies that
used rich controls for prior achievement and family background had not
overestimated the effects of tracking.
These studies are consistent with the conclusion that variation in stu
dents '
experiences in different gro ups and tracks contributes to inequality
in cognitive outcom es. This conclusion rests primarily on a socialization
persp ective , though it relies on allocation and legitimation v iews to explain
wh y the tracks exist and how studen ts are assigned.
racking
expectations and attitudes. Research on the effects of grou p
ing and tracking on non-cognitive outcomes draw s on the same concep tual
framework, hypothesizing that varied socialization experiences lead to
diversity in expectations and attitudes, while leaving explanations of the
sorting process to allocation and legitimation perspectives.
A long tradition of research has established that assignment to college
preparatory programs is associated with higher expectat ions for
post-secondary schooling, compared with other high school experiences
see
Gamoran and B erends 1987 for
a
review). Am ong students with similar
test scores, those in higher tracks are more likely to expect to attend college.
Berends (1992) has found the sam e pattern as early as nin th grade. More
over, Berends observed that amon g stud ents w ith similar test scores and
expectations in eighth grade, those in honors classes are more likely to
maintain high expectations than are students in regular and remedial
classes during the transition to high school.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 81/375
64
Research on att i tud es tow ard s school is m ore am biva lent. Hig h-track
s tudents more of ten conform to the schools ' demands whi le low- track
stu de nt s mo re of ten resis t , bu t these differences m ay be a cau se instea d of
a conse quen ce of track assignm ent (Ga mo ran and Berends
1987).
Wait rowski
et a l. (1983) found no chan ges in delin qu enc y, self-esteem, or at tac hm en t to
school as a consequ ence of ass ignmen t to a col lege-prepara tory pro gram .
H ow ev er , Vanfossen, Jones , an d Spad e (1987) obse rved a w ide nin g
self-esteem gap betw een college- track and oth er s tu de nts , an d B erends ( in
press) found tha t non-col lege s tuden ts reduced the ir eng age m ent in school
and increased the ir d isc ipl ine problems, compared to otherwise s imilar
col lege-bound s tudents .
To the extent th at s trat if ication in schools results in var ied non -cog nit ive
ou tco m es, socializing experiences are said to play a key role. Tea che rs ,
guid anc e counselors , and p eers com mu nicate the m ean ing of di fferent ia l
ass ign m ents , he lp ing to mo t ivate high- t rack s tud ents an d a l ienate those in
low er tracks. Brit ish research ers hav e term ed this the po lar iz atio n
pro ce ss (Lacey 1970; A br ah am 1989).
Cross national corroboration of socialization research.
Research on the
differentiat ion of s tu de nt at t i tu des a nd e xpec tations is as m u ch B ritish as it
is A me rican (G am oran and B erends 1987). Research on ach ieve m ent is
ma inly Am er ican, bu t has been cor robora ted by im por tan t s tud ies in other
cou ntr ies . Several dr aw on Brit ish dat a, inclu ding K erckhoff's (1986) s tu d y
of British secon dary schoo ls , descr ibed abov e. Similar f indings ha ve b een
ob taine d in Israel (e .g., Shavit and Fe ath erm an
1987),
Taiw an (Hsieh 1987),
and Thailand (Lockheed 1987).
Intern ation al research has also sho w n that s trat if ication between schoo ls
can exer t subs tantial effects on achieve me nt gro w th. For exa m ple, K erckhoff
(1986) demonstra ted widening achievement gaps as a consequence of
ass ignment to high-s ta tus grammar schools versus low-s ta tus secondary
m o d em schools in Br itain . G ram m ar school s tud ents a lso gaine d, and
secon dary m od ern s tud ents los t, re lat ive to s imilar s tud ents in co m preh en
s ive schools tha t had heterogeneous s tudent popula t ions . Thus , f indings
for between-school stratif ication mirror those for within-school stratif ica
t ion. Co mp arable resul ts w ere obta ined for the com par ison of academ ic
and vocational high schools in Israel (Shavit and Will iams 1985), and
vocat ional , genera l , and pre-univers i ty secondary schools in the Nether
lands (Eeden 1994). In the United States , specialized public schools have
be en too rare to m ak e such com pariso ns feasible, an d research on pub lic an d
pr ivate schools has not ye t emerged f rom controversy over whether
pr ivate-school achievement a dva ntag es shou ld be a t t r ibuted to the schools,
or to the s t ud en ts w h o atten d pr iva te schools (see Jencks 1985 for a
sum m ary) . W el l -des igned s tudies from other countr ies lend sup po r t to the
conclusion that selective schools ad d value in the sam e w ay that high-status
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 82/375
6
groups and tracks within schools contribute to achievement growth, over
and above initial differences among students.
Stratification as an llocation Process
I argued earlier that work in the socialization tradition implicitly ac
knowledges the value of an allocation perspective for explaining why
different students end up in different schools, tracks, and groups. Yet the
contribution of allocation research is not limited to this indirect impact.
Research emphasizing the allocative properties of school systems has
draw n attention to direct ti s between one's position in the status hierarchy
from one year to the next, and from one level of schooling to the next. (By
direct
ties,
I m ean effects of previous assignm ent on subsequen t assign
ment that are independen t of the presum ed intervening m echanisms, e.g.
achievem ent test scores.) Although this pattern has long been assumed to
hold
(e.g.,
Parsons
1959),
em pirical documen tation is fairly recen t. As of the
early 1980s, Rosenbaum (1976) and Alexander and Cook (1982) had ob
served h ints that track positions tended to persist from year to year, but only
Yogev (1981), in a study of Israeli schools, demonstrated clear, direct ties
between track assignment in one year and track assignment in the next.
More recent work provides substantial support for the early findings.
Gamoran (1989) observed that second grade reading group assignment
depe nd s in part on first grade positions. Pallas et
al.
(1994) also uncovered
connections in reading-group assignment across the elementary school
years.
In similar fashion, Hallinan (1991,1992) has documented the persis
tence of track positions from junior through senior high school.
In a study of a British birth cohort, Kerckhoff (1993) uncovered a high
degree of consistency in status allocations from primary to secondary
school, from secondary to tertiary education, and from schooling to the
labor force. Persisten t ties result not only from skill differences am ong
students found in varied locations, but from long-term allocative effects of
the status positions themselves: Even net of achievement, those in higher
positions in prim ary school typically achieved higher rank at the secondary
level, and those who were well placed in secondary school obtained greater
qualifications th an those less well located, net of earlier test scores. This
pa ttern of direct allocative effects com bined with indirect effects via social
ization (i.e., test score gains) resulted in substantial
cumul tive
effects on life
chances that emerged over the course of the educational career.
In light of these findings, earlier research that attributed track effects on
expectations to differential socialization may need to be reassessed. Instead
of calling on socialization processes such as peer influences to explain how
college-track youth come to hold higher expectations for themselves, a
more parsim onious model w ould state that high school studen ts correctly
perceive that status designations have concrete bearing on their futures,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 83/375
an d they adjust their expe ctat ions accordingly. This rein terpre tat ion is just
w he re Kerckhof f (1976) be ga n alm ost 2 ye ars ago , yet w e still lack sufficient
evidence to determine whether social izat ion is required to account for
effects of tracking on expectations, or whether allocation alone is impli
cated . Stil l, w e hav e ma de great prog ress in identifying the exten t to w hich
allocation pers is ts from year to year an d across levels of the school sy stem ,
and we have solid evidence on the impact of s ta tus al locations.
Institutional ffects of ducational Stratification
Another way of interpret ing outcome differences associated with the
edu cat ion a l s ta tus h ie rarchy is tha t var ied ou tcom es re flec t symbo l ic m ean
ings legitimized in the w ide r society. G am ora n (1986) hy po the siz ed tha t
f irst gra de read ing gro up s a ffect ach ievem ent by enha ncing the m ot iva t ion
of h igh-group s tuden t s and depres s ing engagement among low-group
students ; these processes were seen to occur not through social izat ion per
se , bu t thro ug h anticipatory social ization as a cons eque nce of perc eivin g th e
sym bolic m ea nin g of abil i ty-grou p ass ignm ent; that is, as a self-fulfi ll ing
prop hecy . The evidence , how ever , d id not sup por t theh ypo thes is (Gam oran
1986) . Rather , the assoc ia t ion be tween group rank and achievement was
fully explained by differential instruction, leaving no effects to which
anticipatory social izat ion might be at tr ibuted.
Recently, Pallas et al. (1994) tested Gamoran's (1986) hypothesis with
m or e direct evide nce on cha nge s in s tu de nt s ' expe ctat ions and self-concepts
as a cons eque nce of abil ity gr ou p assign m ent. The au tho rs foun d no
eviden ce of ant ic ipa tory soc ia l iza t ion . H ow eve r , they found su pp ort for
insti tutio na l effects of a different ty pe : A bilit y-g rou p ass ign m en t affected
teachers ' and parents ' expec ta t ions for s tudent performance , ne t of s tu
de nt s ' ac tua l perform ance leve ls . T hus , a l tho ugh the symbol ic m ean ing s of
edu cat ion a l ca tegor ies m ay be obscure to e lem entary school
students
they
may be more apparent and more consequent ia l to parents and teachers .
Since teachers , in par t icular , make recommendat ions and dec is ions about
future placement, effects on teacher expectat ions may have long-term
consequences for s tudents .
Al though ins t i tu t iona l p roces ses do no t accoun t fo r va r ia t ion in
e lemen tary-school achievem ent , they m ay play a grea ter ro le a t the secon d
ary level. G am or an an d Beren ds (1987) ar gu ed th at effects of hi gh sch ool
t racking on achievement may ref lec t shared unders tandings tha t exis t
outs id e schools . T hese no rm s def ine som e s tud ents as col lege-bo und ,
and, perceiving these symbolic designations, s tudents adjust their efforts
accord ingly. N o s tud ies ha ve yet con sidere d the sal ience of inst i tut io nal
proc esses in affecting c ognit ive or s ta tu s outco m es at the seco nda ry level .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 84/375
7
Socialization Allocation and Legitimation as
Complementary Mechanisms
Taken as a whole, this body of work h as mad e substantial progress in
helping us understand the nature and effects of stratification in school
systems. Assign me nt to selective schools, tracks, and ability g rou ps con
tributes to achievement inequality in a variety of contexts. High school
tracking also affects post-secondary exp ectations, and ma y influence m ore
general attitudes tow ards schooling, thoug h evidence on the latter point is
less secure. Researchers have also demonstrated structural linkages from
each point in the system to the next, show ing that one s position in one year
influences placement in the next, independently of performance.
Theoretically, the wo rk indicates that socialization, allocation, and legiti
mation are complementary processes that work simultaneously to bring
abou t these effects. It
is
not necessary to reinterpret socialization resu lts as
allocation effects, as Kerckhoff (1976) suggested; rather, allocation and
socialization processes operate
in concert Aside
from
the
question of changes
in expectations, which cannot be unam biguously attributed
to
either social
ization or allocation, one can trace the workings of each of the empirical
connections described above and interpret them as fundamentally allocative
or socializing processes. The clearest example of this is elem entary school
reading, where allocation is closely linked to previous test scores and
ability-group positions, and learning is tightly connected to instructional
conditions (Barr and Dreeben
1983;
Gam oran 1986; Pallas et a l. 1994).
Nor is it app rop riate, in my view, to consider socialization and allocation
as
special
cases
of legitimation theory , as Meyer
1977)
originally su ggested.
At m ost, this claim is sup por table in the sense that the existence of status
hierarchies within school systems is supp orted by norm ative u nde rstan d
ings that are widely shared. As to the effe ts of stratification, ho wever, most
are more pro per ly und erstood as the consequences of allocation or social
ization. One exception bro ught to light
so
far
is
Pallas et al. s
1994)
finding
that ability-group positions influence teachers and parents expectations
for children in ways that go beyond differences in students actual educa
tional perform ances. Gamo ran and Berends (1987) also hyp othesized that
effects of secondary-school stratification on achievement may result from
institutionalized beliefs over and above instructional conditions, but this
hyp othe sis has no t been tested at the secondary level and ha s failed
to
hold
at the elem entary level (Gam oran 1986; Pallas et al. 1994).
In bringing togethe r socialization, allocation, and legitimation perspec
tives, research in this area has mo ved into new theoretical territory. Recent
writers have taken a longer-term view of the process and effects of educa
tional stratification, evolving a new approach that promises to further
develop and enhance our understanding of schooling and its effects.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 85/375
8
Career Perspective on Educational Stratification
After nearly two decades of research on short-term effects of stratifica
t ion , researchers have begun to cons ider the sys tem as a whole , and i t s
effects on indiv idu als as they traverse the length of the syste m . The
ed uc atio na l career refers to the pa th s an indiv idu al fol lows over t im e from
on e str uc tur al positio n to the nex t (Kerckhoff 1990,1993). In this con cep tion ,
processes of legit imation, al location and social izat ion are closely inter
twined in c rea t ing s t ruc tura l pos i t ions , d is t r ibut ing persons among the
posit ions, and establishing trajectories that tend to pers is t ( though not
invaria bly) over the li fe cou rse. M ore than looking s imp ly at the relat ion
between one 's posi t ion at a given point in t ime, and outcomes that fol low
immedia te ly , the career perspect ive moni tors the progress of individuals
throughout the s trat if ied s tructure of schooling and into the labor market .
In the context of a career perspe ct ive , how eve r , som e points in t ime an d
som e posit io ns in the s tru ctur e m ay be m or e sal ient than oth ers . Bidwell
(1989) has argued that key thresholds , such as the transi t ion from high
school to college or the labor m arke t , de m an d greater at tention . In par t icu
la r , Bidwel l urg es researchers to a t tend to the mul t ip le m ean ing s of edu ca
t ional t rans i t ions . He poin ts to the re la t ion be tw een subs ta nt ive an d
sym bolic me an ing s of t rained capacit ies ; in the lan gu age
ha ve been u s ing,
one could refer to these as the al locative, social izing, and legit imizing
functions of educational s tructure. Successful passage through an educa
t ional t ransi t ion (e .g. , high school graduation) carries meaning in each of
these do m ain s . Hen ce , one may view an educa t ional career as a seque nce
of transi t ions through the s trat if ied educational system.
In a rece nt rev iew , Pallas (1993) explicitly place s schoo ling in the c on text
of the life-course, viewing school participation as one of a series of stages
though which individuals pass as par t of the t rans i t ion f rom youth to
ad ul t ho od . Pa l las show s tha t the s t ruc ture of the educat ional sys tem f igures
pr om ine nt ly in the sequ ence of l ife-course eve nts for ind ivid ua ls . H is
conc lusion is par t ly based on evidenc e that cross-na tional varia t ion in
educational s tructure is t ied to differences in the t iming of important
t rans i t ions . For exam ple , w idesp read availabi li ty of fur ther edu cat io n in
Britain— that is , po st-seco nda ry schooling not at un ivers i ty level an d gen
era l ly pursued s imul taneous ly wi th labor market par t ic ipa t ion—means
that i t is m uc h m or e com m on in Britain than in the U nited S tates for p ers on s
to enter the labor force before com pletin g their edu catio n (see also Kerckhoff
1990).
Research on s trat if ication in school system s seems pri m ed to follow this
approach. Researchers who have been engaged in shor t - te rm analyses of
g ro up ing and t rack ing a re no w p rodu c ing s tud ies of long- te rm, cum ula t ive
effects (Kerckhoff
1993;
H allin an 1994;Hoffer 1994; D au be r 1994). The oreti-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 86/375
9
cally, these s tudies draw on the career perspective, recognizing that a
pers is tent pat tern of al location and social izat ion carries implicat ions that
are more s igni f icant—both subs tant ive ly and symbol ica l ly—than can be
dete cted t hr ou gh analy sis of sho rt- term effects or even of s ingle tra nsi t io ns
in isolat ion. If Kerckho ff 's (1993) analy sis can be gen eral ize d, lon g-ter m
effec ts a re pers is tent and cumula t ive , so tha t what appear to be modes t
influences at a give n po int in t ime turn out to be sub stantia l deflect io ns
over the life course.
An important caution for this l ine of work is that one must avoid an
overly rat iona l view of edu catio nal organ ization and al location. For the
sake of par s im on y, there i s a tendency to assum e tha t ass ignm ent to schools
and c lasses i s an order ly , logica l process , in which s tudents ' presumed
ed uca tion al nee ds are taken into accou nt. This is no t to say inequali ty is
ignored - on the contrary, a variety of biases influence the ass ignment
process , and these ha ve been examin ed c lose ly— but the assu m pt io n is tha t
ass ignment occurs in an orderly fashion.
Ca se s tudies of schoo ls sho w this view is far too s im plis t ic . Part icu larly
a t the secondary leve l, s tud ent a ss ignm ent i s m ore a m at ter of sch edu l ing
an d logis t ics tha n an issue of edu catio nal ne ed s (DeLany 1991). Ev en at the
e lem entary leve l , there is en orm ou s over lap in s tud ent abi l i ties am on g
thos e ass igned to gr ou ps of differing s tatu s , an d i t is not clear w hy so m uc h
ov erl ap exists (Pallas et al. 1994). O ne possib ility is tha t over tim e, the
variance in s tudent performance increases within each s tatus level (e .g. ,
w ith in high g ro up s, low g ro up s, ho no rs classes , reg ular classes , e tc .) . Yet
s tudents ' pos i t ions in the h ierarchy tend to pers is t despi te the i r var ied
perfo rm anc es. A lth ou gh som e shifts in ran k occur, these are not alw ay s at
the m arg ins , beca use of sche dulin g and other logist ical con straints (Ha ll inan
and Sorensen
1983;
DeLany 1991). Con sequ ent ly , over lap increases am on g
students ass igned to different levels .
This interpretat ion could be tes ted by asking whether the degree of
ove r lap wid en s as s tud ents proceed thro ugh the school sys tem. Is there
m ore over lap in achievem ent am on g s ix th grad e than second grad e rea ding
gro up s? O ne m igh t a lso assum e tha t key t rans i tions (such as the t rans i t ion
to high school) are used to real locate s tudents to reduce overlap, but case
studies make one skeptical that this occurs (DeLany
1991;
G am ora n 1992) .
For the career persp ectiv e on s trat ificat ion in schools , this cau tion hig hlig hts
the importance of considering the organizational context of s trat if icat ion
along s ide the progre ss of individu al careers .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 87/375
7
Future Dir ections for Research on Educational areers
By the end of this de cad e, w e will ha ve m uc h mo re information ab ou t the
long -term effects of stratif ication in schools , inclu ding m ore se cure kn ow l
ed ge ab ou t the al locative role of schoo ling ov er an exten ded pe r io d, an d i ts
cum ulat iv e social iz ing impact . This m uc h can be seen in wo rk cur ren t ly
un de rw ay . Beyond tha t , mo re w ork is nee ded .
First , des pite the progres s m ad e, w e ne ed s t i ll better inform ation ab ou t
the m ec ha nis m s thro ug h whic h the effects of s trat ification o ccur . A lth ou gh
evidence about the overall effects of stratif ication comes from a variety of
coun tr ies , evidence on mec hanism s is exclus ively Am er ican. For exam ple ,
outs ide the
U.S.,
d o effects of m id dl e school ability g ro up in g reflect ins tru c
t ional differences? I t is possible that in other countr ies w he re gr ou p
ass ign m ent is m ore formalized, effects of gr o up assig nm ent reflect s ym
bolic s tatus dis t inctions—i.e. , a l location and legit imation—more t i tan in
stru ctio na l or oth er socialization effects a s in the Un ited
S ta tes.
H ence , there
is a need for comparative analysis of the mediating role of instruction in
pr od uc in g effects of selective schools , classes , an d gr ou ps .
Second, a l though much is known about how educat ional s t ra t i f ica t ion
affects s tu de nts betw een the ages of
a nd 18, w e kno w m uch less abo ut la ter
and especially ear l ier per iods. Yet the expansion of both pre-school and
pos t- secondary educat ion has coincided with increas ing inequal i ty of ac
cess to the typ es of schooling with th e hig hes t s tatus (nurse ry schoo ls at th e
pre-sch ool level , and 4-year colleges at the po st-sec on dary level) (A lsalam
et a l . 1992). Pre-kind ergar ten m ay be the m ost s tra ti fied d om ain of educ a
t ion at the present t ime, if one takes into account al l the var ious types of
extern al care for yo un g children. Because there is m uc h less pub lic fun din g
for education for pre-school-aged children compared with funding af ter
age 5 (Head Star t is the notable exception) , family wealth is a s trong
pre dict or of the type of care an d edu cation children receive. Do es s trat if i
cation pr ior to kin de rga r ten reinforce or redu ce inequ ali ty in read ines s for
school? O n the one ha nd , if pre-scho ol is un eq ually al located, for exa m ple
alon g socioeconomic l ines, then inequ ali ty may be reinforced. On th e other
hand, if var iat ion in pre-school care is less unequal than var iat ion in
chi ldre n s ho m e l ives , then dispar i t ies in school readiness m ay be redu ced ,
de spit e the s trat ified na tu re of pre-scho ol care. M ore research has be en
co nd uc ted on stratification at the po st-sec on dary level (e.g. , Brint an d
Karabel 1990), bu t this do m ain re m ains less scrutinize d th an K-12 ed uc a
tion, yet it is hig hly stratif ied an d dese rve s m ore atten tion from rese arch ers
concerned with s trat if ication and inequali ty.
Third , mo re research is ne ed ed on the transit ion out of ed uc atio n an d into
the workforce. Studies of cum ulat ive effects of gr ou pi ng an d trackin g in the
U.S. are l imited to the edu cation al system an d do not m ov e be yo nd i t. In the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 88/375
71
future i t wil l b e essen tial to exa m ine the im pact of the s trat if ied s tru ctu re of
schooling on l ife chances , taking into account not only educational oppor
tuni t ies bu t labor ma rke t par t ic ipa t ion , as a consequence of the exper iences
of scho oling . For the British case, Kerckhoff (1993) ha s sh o w n that ed uc a
t iona l careers ha ve long - te rm consequ ences .
Fo urth , sociologis ts w ork ing in this area could w rite m or e explici tly for
pol icy-makers . Sociologica l research in th is domain has made important
policy contrib utio ns: Evidence on the sal ience of with in-scho ol variat io n in
s tudent achievement , on the assoc ia t ion of s tudent achievement wi th
soc io dem ogra phic co ndi t ions tha t vary w i th in schools, and o n the l inkages
be tw ee n s trat ificat ion in schools an d s trat if ication in the w ide r society, a l l
der ive from wo rk by sociologis ts . M ore explicit co m m en tar y on these
findings for general audiences would raise the quali ty of discourse about
such issues . A recent exam ple is Ha l l inan ' s exchange wi th Oakes in
Sociology of Education (1994). I t is im po rtan t that sociologis ts brin g their
evidence and ins ights more d i rec t ly to the pol icy audience , in ways tha t
ad dre ss the choices tha t confront ed uca t ion a l d ec is ion-mak ers .
otes
This paper was written at the Consortium for Policy Research in Education
(CPRE), Wisconsin Center for Education Research, University of Wisconsin-Madison.
CPRE is supported by the Office of Educational Research and Improvement, U.S.
Dep artmen t of Education (grant
no.
OERI-R117G10007). Findings and conclusions
expressed in this paper are those of the author and do not necessarily reflect the
views of the supp orting agencies.
References
Abraham ,John. 1989. Testing Hargreaves'and Lacey's D ifferentiation-Polarisation
Hyp othesis in a Setted Com prehensive. British journal o f Sociology
40:46-81.
Alexan der, Karl L., and M artha A. Cook. 1982. Curricula and Coursew ork: A
Surprise Ending to a Familiar Story. American Sociological Review 47:626-640.
Alexander, Karl L., M artha A. Cook, and Ed wa rd L. McDill. 1978. Cu rricu lum
Tracking and Educational Stratification. American Sociological Review 43:47-66.
Alexander, Karl L., and Edward L. McDill. 1976. Selection and Allocation within
Schools: Some Causes and Consequences of Curriculum Placement. American
Sociological Review 41:963-980.
Alsalam, N abeel, Laurence
T
Ogle, Gayle
T
Rogers, and T hom as M. Sm ith. 1992.
The
Condition
of Education 1992. W ashington, DC: U .S. Departm ent of Education.
Averch, H.
A., S
J
Carroll,
T
S
Donaldson,
H.
J
Kiesling, and
J
Pincus. 1972. How
Effective is Schooling
A
Critical Review and Synthesis
o f
Research Findings. Santa
Monica, CA: RAN D.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 89/375
72
Barr, Rebecca, an d Robe rt Dreeb en. 1983. How
Schools
Work Chicago: Univers i ty
of Ch icago P ress.
Beren ds, Mark. 1992. Effects on School Or ientat ions and Ac hievem ent in the
Transi t ion f rom Middle to High School . Un publ i she d Ph .D. thes i s . D epa r tm ent
of Sociology, Universi ty of W isconsin-M adison.
Bere nds, M ark. In press. Edu cat ional Strati f icat ion and Stude nts ' Social Bon ding
to School . British
Journal
of
Sociology
of Educa tion.
Bidwell , Ch arles E. 1989. The M eanin g of Edu cat ional A ttainm ent .
Research in
Sociology
of Education and Socialization. Vol. 8 1989:
Selected Methodological
Issues
edi ted by Krishnan Nam bood i r i and Ronald Corw in . G reenwich ,
C T :
JA Press.
Brint, Stev en, an d Jer om e Karabel. 1989.
The Diverted Dream: Comm unity ollegesand
the Prom ise of Educational Opportunity in America 1900-1985. New York: Oxford
Universi ty Press.
Co lem an, J., E. Cam pbell , C. Hobso n, J M cPart land , A. M ood , F. W einfield , and R.
York. 1966. Equality of Educational Opportunity. W ashington , DC : U.S.
Government Print ing Office.
Da uber , Susan. 1994. Track ing and Transi t ions thro ug h the M iddle G rad es:
Ch ann el ing Educat ional Trajectories. Pap er pres ented at the an nu al m eet in g of
the American Sociological Association, Los Angeles.
D eL any , Brian. 1991. Alloca tion, Cho ice, and Stratification w ithin Hig h Schoo ls:
H ow the Sort ing Machine Co pes .
American Journal of Education
99:181-207.
Eed en, Pieter van den . 1994. Edu cat ional Select ivi ty from the Mu lt ilevel Perspec
t ive .
Unpubl i shed manuscr ip t , Depar tment o f Socia l Research Methodology ,
Vrije Univers i ta t , Am sterdam , the N ether land s .
G am or an , A da m . 1986. Instruct ional and Inst i tut ional Effects of Abil i ty G ro up
ing .
Sociology of Education
59:185-198.
G am oran , Ad am . 1987. The St ra ti fica tion of High School Le arn in gO ppo rrun i t ies .
Sociology of Education 60:135-155.
G am oran , Ad am . 1989. M easur ing Cu rr icu lu m Differen t ia t ion . American
Journal
of Education
97:129-143.
G am ora n, A dam . 1992. Access to Excellence: A ssignm ent to H on ors English
Classes in the Transi t ion from Middle To High School .
Educational Evaluation
and Policy Analysis 14:85-204.
G am oran , A dam . 1993 . Al te rna t ive Uses of Abi l ity G rou ping in Secon dary
Schools: Can W e Bring High-Q uali ty Instruct ion to Low-A bil ity Classes?
Ameri
can Journal of Education
101:1-22.
G am or an , A da m , and M ark Bere nds. 1987. The Effects of Stratification in
Sec onda ry Schools: Synthesis of Surv ey and Ethn ogra phic Res earch.
Review of
Educational Research
57:415-435.
G am ora n, Ada m , and Robert D. M are. 1989. Seco ndary School Strati f icat ion and
Educat ional Inequal i ty: Compensat ion, Reinforcement , or Neutral i ty?
Ameri
can Journal of Sociology 94:1146-1183.
G am oran , Adam , Mart in Nyst rand , M ark Berends , and Paui l C. LePore . In press .
An Organizat ional Analysis of the Effects of Abil i ty Grouping.
American
Educational Research Journal.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 90/375
7
H all inan , M aure en T. 1991. School Differences in Tracking Structures an d Track
Ass ignmen t s . Journal of
Research
on Adolescence 1:251-275.
H all inan , M aur een T. 1992. The Or gan izat ion of Stu den ts for Instru ct ion in the
Middle School .
Sociology of Education
65:114-127.
H all inan , M au ree n T. 1994. Effects of Track ing on Ac hievem ent of Black and W hite
Stude nts : A Lon gi tud ina l Study . Pape r p resen ted a t the ann ual meet ing of the
American Sociological Associat ion, Los Angeles.
H all inan, M aur een T., an d Aage B. Sorense n. 1983. Th e For m ation and Stabi l ity
of Ins t ruc t iona l Groups .
American Sociological Review
48:838-851.
H ey ns , Barb ara. 1974. Social Selection an d Stratification w ithi n Schools.
American
Journal of Sociology 79:1434-1451.
Hoffer, Th om as. 1992. M iddle School Abil i ty G rou pin g and Stu den t Achiev em ent
in Science an d
Ma thematics. Educational Evaluation and Policy Analysis
14:205-228.
Hoffer, Th om as. 1994. Cu m ulat iv e Effects of Cu rric ulu m Trac king on Stu de nt
Ac hievem ent . Pap er p resen ted a t the ann ual mee t ing of the Am erican Socio
logical Assoc iat ion, Los An geles.
Hoffer, Th om as, an d A da m G am ora n. 1993. Effects of Instruc t ional Differences
am on g Abi l i ty G rou ps in Middle-School Sc ience and Mathem at ics . Paper
presented at the annual meet ing of the American Sociological Associat ion,
Miami .
Hs ieh, H-C . 1987. Abil i ty Strati f icat ion in U rban Taiwa nese Sec ond ary Schools.
Bulletin of the Institute of Ethno logy
64:205-252.
Jencks , Chri s top her L. 1985. H ow Much Do H igh School Studen ts Learn?
Sociology of Education 58:128-135.
Jencks, C. L., M . S. Sm ith, H . Aclan d, M . J. Bane, D. Co hen , H . Gintis, B. Heyns , and
S. M ichaelson.
Inequality.
N ew Yo rk: Basic Books.
erckhoff A lan C. 1976. The Status A ttainm ent Process: Social izat ion or Alloca
tion?
Social Forces
55:368-381.
erckhoff A lan C. 1986. Effects of Ab ility G ro up in g in British Se co nda ry Schoo ls.
American Sociological Review
51:842-858.
erckhoff A lan C. 1990.
Getting
Started:
Transition to Adulthood in G reat B ritain.
Boulder , CO : W estv iew.
erckhoff Alan C. 1993. Divergent Pathways: Social Structure and areer Deflections.
Cambridge: Cambridge Univers i ty Press .
Lacey, Colin 1970. Hightown Grammar. Manchester: Manchester Universi ty Press.
Lockheed, Marlane E. 1987.
School and lassroom Effects on Student Learning Gains:
The Case of Thailand
World Bank Educat ion and Training Series, Discussion
Pape r N o. EDT 98 .
M eye r, Joh n 1977. Th e Effects of Ed uca tion as an Ins titutio n. American
Journal of
Sociology
83:55-77.
Mostel ler, Fredrick, and Daniel P. Moynihan, Editors. 1972.
On Equality of Educa-
tional Opportunity. N ew York: Vintage.
M ur ph y, Joseph, and Phil ip Hall inger. 1989. Equity as Access to Lear ning:
Curricular and Instruct ional Treatment Differences.
Journal of Curriculum
Studies
21:129-149.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 91/375
74
Oakes, Jeannie, Adam Gam oran, and Reba N. Page. 1992. Curriculum Differen
tiation: Opportun ities, Outcomes, and Mean ings. Pp. 570-608 in
Handbook of
Research on
Curriculum edited by Philip
W
Jackson. Washington, DC: American
Educational Research Association.
Pallas, Aaron M.
1993.
Schooling in theC ourse of Human Lives: The Social Context
of Education and the Transition to Adulthood in Industrial Society.
Review of
Educational Research
63:409-447.
Pallas, Aaron M., Doris E. Entwistle, Karl L. Alexander, and M. F. Stluka. 1994.
Ability-Group Effects: Instructional, Social, or Institutional? Sociology of Edu-
cation
67:27 46.
Parsons, Talcott 1959. The School Class as a Social System: Some of its Functions
in Am erican Society.
Harvard Education Review
29:297-313.
Rosenbaum, James E. 1976.
Making
Inequality:
The Hidden Curriculum of High School
Tracking. New York: Wiley.
Row an, Brian, and A ndrew W Miracle, Jr 1983. Systems of Ability Grouping and
the Stratification of Achievem ent in Elementary Schools.
Sociology of Education
56:133-144.
Sebring, Penny A. 1987. Consequences of Differential Amounts of High School
Coursework: Will the New G raduation Requirements
Help? Educational Evalu-
ation and Policy nalysis
3:258-273.
Shavit, Yossi, and D avid L. Featherman. 1988. Schooling, Tracking, and T eenage
Intelligence.
Sociology o f Education
61:42-51.
Shavit, Yossi, and Richard Williams. 1985. Ability Group ing and C ontextua l
Determinants of Educational Expectations in Israel.
merican Sociological
Re-
view 50:62-73.
Slavin, Robert E. 1990. Achievem ent Effects of Ability Grouping in Secondary
Schools: A Best-Evidence S ynthesis.
Review
of
Educational Research
60:471-499.
Sociology of Education. 1994. Exchange. (Contributions by M. T. Ha llinan and
J. Oakes.)
Sociology
o f
Education
67:79-91.
Sorokin, Pitirim A. 1927.
Social
M obility. New York: Harper.
Vanfossen, Beth E., ames D.
Jones,
and Joan Z. Spade. 1987. Curriculum Tracking
and Status Maintenance.
Sociology of Education
60:104-122.
W aitrowski, Michael
D.,
Stephen Hansell, Charles R
Massey,
and Donald
L.
Wilson.
1982.
Curriculum Tracking and Delinquency.
merican Sociological Review
47:151-160.
Yogev, Avraham. 1981. Determ inants of Early Educational Career in Israel:
Further Evidence for the Sponsorship Thesis.
Sociology of Education
54:181-195.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 92/375
Educational Tracking
during the Early Years:
First G rade Placements
and Middle School Constraints
Karl
L
lexander
Doris R Entwisle
The Johns Hopkins Univerisity
In l ine w ith the early educ ation al s trat ification pers pec tive, in w hich the
school- to-work t rans i tion an d pers istence in the educat ional sys tem thro ugh
and beyond high school were the predominant concerns (e .g . , Hauser ,
1970), research on tracking thus far has focused mainly on curriculum
different ia t ion a t the upper grade leve ls , and then mainly on individual
a t ta inm ent ou tcom es , l ike test scores , m arks and aspi ra t ions . Tracking
sys tem s as s t ruc tura l cons t ra in ts on the student career are jus t beg inn ing to
co m m an d at ten tion. Research on track m obil i ty at any level of scho oling is
spars e ; sparser s t il l a re s tudies tha t exam ine placem ent pa t te rn s
across
levels
of scho olin g. Kerck hof f s (1993) Diverging Pathways is the mo s t c om prehe n
sive trea tm en t of thes e issues pres ently available. Trac king a coh ort of
Bri t i sh youth f rom e lementary school in to young adul thood, Kerckhoff ' s
analy sis sh ow s tha t t rack plac em en ts an d other s tru ctu ral features of school
organiza t ion cons t ra in la te r p lacements and performance outcomes a l l
throughout the school ing process , f rom the ear ly pr imary grades in to the
pos tse con dary level . Recent ly , severa l ana lyses ha ve be gu n to addr ess
these mat ters in the
U.S.
too (G am ora n, 1992; H allin an , 1992; Hoffer, 1992;
Stev enso n, Schiller and Schn eider, 1994), b u t al l focus on track ing at the
upper grade levels , in middle school and beyond. This neglect of early
tracking in
U.S.
research is un fort un ate, as edu catio nal sort in g an d select
ing in the ear ly pr im ary gr ades no do ub t se ts the s tage for m uc h tha t
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 93/375
7
follows (e.g., En twisle an d A lexa nd er, 1993).
The present paper descr ibes t racking pa t te rns in f i rs t grade , when the
process commences , and how they ar t icula te wi th in i t ia l program place
m en ts in m idd le school. It thus begins to explore s t ruc tura l cons t ra in ts tha t
orig inate in child ren 's very f irst enc ou nter s w ith edu catio nal t racking . The
da ta co m e from ou r Be ginning School Stu dy (BSS), w hich s ince 1982 has
been moni tor ing the academic progress and personal development of 790
youngsters who began firs t grade that year in 20 Balt imore City public
schoo ls . The sam plin g of schools spa ns the ran ge of socioecono m ic levels
in the City system, as well as different school integrat ion contexts , with
w h i t e s , r e l a t i v e l y w e l l - t o -d o n e i g h b o rh o o d s a n d i n t e g ra t e d s c h o o l s
ove rsam pled to sus ta in s t ra tegic com parison s . W ithin schools s tud en ts
w ere selected ran do m ly from 1981-82 kin der gar ten rosters , w ith su pp le
m en tal sam plin g in the fall to pick u p new en tran ts . All f irst gr ad e
classro om s are cove red in the f inal be gin nin g school coho rt . The project
n ow is in i ts twelfth yea r. Soon we wil l kn ow h ow m an y of the s t ud y
yo un gste rs are on-t im e twelfth g rad ers , bu t almo st certainly i t w il l be w ell
u n d e r half Ac cord ing to interview d ata o btain ed from 82% of the original
790 in the spring of Year 11, jus t 48% at tha t poin t w ere e leventh grad ers .
Fifteen per cen t (N = 99) ha d already d ro pp ed out and a no ther th ird w ere a
yea r or mo re be hin d (22.4% in tenth gr ad e; 12.4% in ni nt h g rad e; one s t il l in
e ighth grade) .
1
The se yo un gste rs ail s tarted o ut together as firs t gr ad er s in
the fall of 1982, bu t ove r the years mu ch h as ha pp en ed to m ov e the m onto
different edu cat ional pa thw ays .
2
It s reasonable to suspect tha t edu cat ion al
tracking has played a role .
The Several im en sion s of Early Tracking
A lthou gh t racking comes in m an y guises,
3
research ten ds to ad dre ss the
issue piecem eal. Typically a s ingle form of tracking , l ike abil i ty g ro up ing ,
is exam ined in i so la t ion f rom th e o thers . Indeed , some of the m ore com m on
forms of t racking in the pr ima ry g rades usual ly a re n ' t even tho ug ht of as
such -retention an d Special Edu cation placem ents , for exa m ple. Fo llow ing
Oakes (1992), and before her Sorensen (1970; 1987), we favor a broad
con struction of tracking as inh erin g in the org aniza tiona l differentiat ion or
s tud ents . Abi li ty gro up ing in the pr im ary g rade s , curr icular d ist inc t ions in
the middle grades and high school , Specia l Educat ion programs and
re tent ion a l l regroup s tudents , reduce he terogenei ty in the ins t ruc t ional
unit (the goal of virtually all systems of tracking), and confer distinctive
org aniz ation al identi t ies . Abil i ty gro up ing is the form of early tracking
m ost often s tudied from an edu cation al s trat if ication per spec tive, b ut th ere
are im po rta nt paral lels that cut across al l these forms of tracking . Like
assignment to low instructional groups, retention and Special Education
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 94/375
placements usually are based on assessed abil i ty or correlates thereof
Given the achievement-or iented ideology tha t pervades the educat ional
sys tem , they a lmost certa inly car ry negat ive conno ta t ion. Rep eaters are
sepa ra ted f rom th eir same-age peers and de pa r t from the norm al t im etable
of grade progressions. Special Education typically occurs outside regular
classes and also often has the effect of throwing children off schedule. This
asynchrony makes repeaters and Specia l Educat ion s tudents conspicuous
and thu s vulnerab le to negat ive labell ing.
Reading Groups in First G rade
Acco rding to M cPar t land, Coldiron and B raddock
(1987),
more th an 90%
of elem enta ry schoo ls use with in class abil ity grou pi ng for readi ng in f irs t
grad e . W ithin c lass gro upin g (as dis tinc t f rom betw een c lass grou ping )
benefits pupils of some or all ability levels (Kulik and Kulik, 1987; Slavin,
1987),
b u t the co nd it ion s that al low posit ive effects to m ater ial iz e an d
whether benefits occur for both reading and math are disputed Archam-
bault, 1989).
4
In review ing research on gro up ing in the primary gra de s, for
example, Slavin (1987) found too few methodologically sound studies to
pe rm it con clusions for read ing , bu t s ince rea din g is the area of the c urr icu
lum that relies m ost of ten on small gro up instruction this leaves the m atte r
ve ry m uc h open (on ly about
2 5 %
of schools use small gr ou ps in ma th - see
McPar t land, Coldiron and Braddock, 1987) . Beyond the ques t ion of how
grouping af fec ts per formance, where both sent iment and evidence are
mix ed, there a lso is concern abou t the s t igma tha t su ppo sedly a t taches to
low g ro up m em be rsh ip (e.g. , Rist, 1970).
Smal l ins t ruct ional gro up s help teachers deal wi th prac t ica l pro ble m s of
classroo m m an ag em en t. Even in schools that gro up entire classes by abil i ty
or readiness tests , fur ther grouping of ten occurs inside the classroom.
Splitting a class of 30 into three or four groups of about the same level of
read ines s simplif ies lesson pl an nin g and m ak es i t easier to m eet ind ivid ua l
s tu de nts ' ne eds . G roup ing , in th is sense , i s an organiza t ion al response to
an organiza t ional pro ble m (Dreeben, 1984:83). In conseq uence , gro up ing
arrangements of ten are broadly s imilar f rom place to place regardless of
school context or s tudent character is t ics , l ike the dis tr ibution of abil i ty
levels in the class (Hallinan and Sorensen, 1983).
As an organ iza t ional imp era t ive , ins t ruct ional gro up ing crea tes bot to m
d o g gr ou ps ma d e up of a third to a fourth of the class w he rev er it is use d.
This he lps expla in w hy conven t ional social and dem og raph ic r isk fac tors
(i.e.,
family SES level; race/ethnicity), ordinar i ly powerful predictors of
educat ional outcomes , have prac t ica l ly no bear ing on group placements
(Haller , 1985; H aller a nd Da vis , 1980; Pallas , Entw isle, A lexa nd er an d
Stluka, 1994; Sare nse n an d H all inan , 1984).
5
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 95/375
8
Grade Retention
Th ere are no goo d nationa l data on grad e retention, bu t reten tion rates in
m an y locali ties are qui te high. According to da ta comp iled by Sh epard and
Smith (1989) for 14 states for the 1985-86 school year, grade-specific reten
t ion rates in the 7% - 8% rang e are com m on . These data are con sis tent w ith
off-time rate s at the en d the elem en tary y ears of fifty pe rce nt or m or e, w hic h
nat ional da ta show in many urban areas wi th large pover ty level and
minority enrollments (Bianchi, 1984) . In Shepard and Smith 's data, reten
tion rates consistently are highest in first grade, averaging 11% for the 13
state s w ith first gra de da ta. Eight of the 13 states ha d first gra de rete ntio n
rates above 10%. By way of comparison, the
highest
f igure in grad es tw o
th ro ug h six wa s a li tt le over
8 .
In Balt imore too retentio n at the elem enta ry
level is hig he st in first gra de . Both the 1982
BSS
da ta and sys tem -wide da ta
for the 1989-90 scho ol yea r (Kelly, 1989) p u t the first gra de rate at o ver 16% .
For som e time sent iment in the research com mu nity reg ardin g re tent ion
has been a lm ostun iversa l lyn egat ive . In 1971,for exam ple , Ab idin , Go l laday
an d H ow erto n referred to retention as an unjustif iable, discr imin atory a nd
noxio us edu cat iona l pol icy, the
99
me et ing of the Am er ican Edu cat ional
Research Associa tion held a sym po sium ent i t led Retent ion: Processes and
Co nse qu enc es of a M isgu ided Practice, an d qu otes l ike the following
ab ou nd (Ho use, 1989:210); .. .. the evid enc e is extensive and un equ ivo cal.
I t inclu des test scores , surv eys , perso nali ty an d emo tional adju stm ent
measures, case s tudies—everything from elaborate s tat is t ical analyses to
asking s tude nts ho w they feel. Alm ost every thing points in the sam e
direction—retention is an extremely harmful practice.
Such swe epin g prono un cem ents a lmost certa inly overs ta te the case for
the kinds of outcomes most of ten considered in evaluations of retention
(e.g. , m ark s, test scores and att i tud es) . Studies do not w eigh in c onsis ten tly
against retention (see, for example, Pierson and Connell , 1992; Peterson,
DeG racie, an d A yab e, 1987; Re yno lds, 1992), an d ou r ow n asse ssm ent
(Alexander , Entwisle and Dauber , 1994) over the elementary and middle
schoo l years f inds m ainly posit ive, no t negativ e, effects of s ingle retention s
for all these criteria.
6
But w ha t of retention as a s tructu ral constrain t on
ch ildr en 's later schooling? De spite concerns that rep eatin g a gra de closes
off o pp ortu nitie s later, evid ence on the issue is alm ost entirely circu m stantia l.
Ear ly re tent ion cons is tent ly predic ts la ter dropout and most dropouts
are in non-academic tracks (Cairns, Cairns and Neckerman, 1989; Fine,
1991; Stroup and Robins, 1972; Lloyd, 1978), so there certainly are indica
t ions of retentio n's relevance for tracking patte rns across the s tud en t career.
H ow ev er , the only direct evide nce of such l inkages comes from the NELS-
88 da ta. Stev ens on , Schiller an d Sch neid er (1992) find, for ex am ple , that
elementary school repeaters are less l ikely than other youngsters to take
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 96/375
79
high level courses in middle school and high school, even after adjusting for
backgro und factors and test scores. This patte rn is consistent w ith the
mobility constraint
idea;
unfortunately, the
NELS 88
data rely on retrospec
tive reports of children's prior promotion histories and this particular
analysis also used self-reports to determine both middle school and high
school curricu lum placements. Thus, while early retention probably limits
curricular options la ter, so far there is little solid evidence to this effect.
Special ducation
From the m id-seventies to the early nineties the num ber of children of all
ages receiving Special Education services increased from 3.7 million to 4.8
million, or from about 8% to just over 12% of 3 to 21 year-olds (Office of
Special Education Programs,
1991;
Viadero, 1992). Moreover, despite P.L.
94 142 and the mainstreaming movemen t, most Special Education stude nts
continue to receive services outside regular classrooms (Office of Special
Educa tion, 1991). Nevertheless, with few exceptions (e.g., Carrier, 1986;
Mehan, 1992; Mehan, Hetweck and Meihls, 1985; Mercer, 1974) Special
Education still has received little attention as an educational track.
According to the Office of Special Education Programs
(1991),
between
1976 77
and
1989 90
students classified as learning disabled increased from
24.9%
of the total receiving services to
50.0%,
by far the largest percentage
increase registered
in
any of
the
standard placement categories
(e.g.,
m ental
retardation, serious emotional disturbance, speech or language impair
ments, hearing impairments, etc.). Learning disabilities are defined by a
marked discrepancy between ability and performance, but often the cat
egory is a catch-all, with screening frequently done simply through test
scores
Singer,
Palfrey, Butler and W alker,
1989).
Criteria for d istingu ishing
between low-achieving and learning disabled children are ambiguo us and
practices vary greatly from place to place (e.g., Algozzine and Ysseldyke,
1983;
Reynolds, 1984). Minority youngsters are greatly overrepresented
among the educationally mentally retarded and in some other Special
Education categories, raising equity questions that seem invariably to arise
in the context of tracking (e.g., Heller, Holtzm an and Messick, 1982).
As with ability grouping and retention, assessments of traditional Spe
cial Education prog ram s tend to be critical (e.g., M adden and Slavin, 1983;
Leinhardt and Pallay,
1982). Again,
however, practically noth ing
is
known
about how Special Education placem ents tie in with other forms of tracking.
This hold s bo th for short-term
issues,
like how placem ents ove rlap, and for
longer term issues, like how early p lacements limit oppo rtunities later.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 97/375
8
Ear ly Tracking in
th
BSS
First Grade Placemen ts
As shown in Table 5 .1 , just ov er 16% of the BSS coh ort w as h eld b ack at
the en d of first gra de , a figure in line with c ity-w ide statistics; 13% received
Special Education services in their f irst
or
second school years;
7
a nd wh e n
child ren are classif ied accord ing to w he the r they w ere in the low est g ro up
in their classroom , in the hig hes t gro up , or in an inte rme diate level gro up ,
8
22% were located in the lowest .
9
In the r ightmost colum ns , low and high
pla cem en ts are tal lied across these three facets of ear ly track ing. M ost
youngs ters (a lmost 70 ) are spared low placemen ts a l together , bu t s izeable
m ino ritie s are placed low in on e (17%), tw o (9%), an d all thre e (5%) area s.
The se low plac em ents are our m ain concern. The l i terature on abil i ty
gro up ing , re tent ion a nd Special Edu cat ion tend to t rea t them as isola ted in
chi ld ren 's exper ience . In the rea l wo r ld , thoug h, they of ten com bine . F igure
5.1 sho w s how they ov erlap in the case of firs t grad e repea ters . Th ere are 88
such yo un gste rs in the BSS w ho se Special Edu cation s ta nd ing an d read ing
gro up level a lso are kn ow n. Seventy-seven percent were in low read ing
gr ou ps , 45 % of w h om also received Special Edu cation services (equ iva-
lently, a third of the 88 were assigned to Special Education
and
in low
read ing grou ps) . The cor resp ond ing f igures for chi ldren prom ote d a t the
en d of first g ra de w ere in the 10% -1 2% rang e (see lower c har t in Figure 5.1) .
Just a fifth of the retainees were spa red bo th other low pla cem ent s vers us
m ore tha n 80% of the pro m ote d gro up .
Overall , then, almost 80% of f irs t grade repeaters had multiple low
placem ents . So too did 61 % of childre n in low rea din g gro up s and 56% of
tho se receiving Special Edu cation services . The specific configu rations are
prese nted in F igure 5 .2 (where high placem ent pa t tern s a lso are display ed) .
H ere Lo w m ea ns held back , receiving Special Edu cation services an d
be ing in the lowes t r ead ing grou p . Conver se ly , h igh m ean s p rom oted ,
not receiving Special Education services and assignment to the class 's
highes t reading gro up . Retent ion and Special Educat ion are dicho tom ies ,
so lo w a lso m ean s not h igh . In other w ord s , reading gro up level i s w ha t
dis t inguishes the two classif ications.
O ve r half of the 3 0% plac ed low in first gra de (194 of the 630 w ith d ata on
al l three measures) had jus t one low placement—26% in low reading
gro up s, 10% retained a nd 19% receiving Special Ed ucation services . Th e
three in terv ent io ns in these ins tances are separable , wh ich is ho w they are
treate d in mos t of the literatu re— on e facet of tracking in isolatio n from the
othe rs . H ow ev er , abo ut 14% of the total and just und er half of those w ith
low placem ents occupied two or three low t rack s lots . The m ain pa t tern
combines low reading group with grade re tent ion, but there a lso are
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 98/375
81
TABLE
5 1
First Grade Track Placements: Reading Group Level, Special
Education, Retention, and Tallies of Low and H igh Placements
1.
Reading Group
Lowest
In t e rmed ia t e
Highes t
2. Special Education
Yes
N o
3 .
Retained
Yes
N o
2 2 .1%
(132)
38 . 3%
(228)
39.6%
(236)
13.3%
(105)
86.7%
(685)
16.2%
(128)
83 .8%
(662)
#
Low Placements
0
1
2
3
69.
2 %
(436)
17.0%
(107)
9.2%
(58)
4.6%
(29)
# H i # t
0
1
2
3
Placements*
5. 6%
(35)
15.4%
(97)
4 3 . 3%
(273)
35.7%
(225)
Includes lowest reading group, retained and receiving Special Education ser
vices.
In clud es highest reading grou p, not retained and not receiving Special Educa
tion services.
promoted chi ldren who were in low reading groups and received Specia l
Edu cat ion services and others wh o were in Specia l Edu cat ion an d held back
but not in low reading groups . The numbers involved are smal l in these
ins tances , but the pat terns we assume are re l iable , so sys tem-wide in the
BCPS, an d nat ion -w ide in places l ike Balt imore, a great m an y ch ildren are
tou che d by all these experiences, s ingly and in com bina tion. Inde ed , alm ost
15% are tracked low in al l three areas: placed in a low reading group,
assigned to Special Education and held back at year 's end.
These pa t tern s br in g to l ightso m e of the complexi ty of ear ly t racking. For
examp le , p rob lem s as soc ia ted wi th low t r ack p lacements might co m po un d
as the n um be r of suc h p lacem ents increases. If this is the case, then researc h
on re tent ion, or on group ing, or on Specia l Edu cat ion wil l un de rs ta te ho w
tracking per
s
im pac ts s tud en ts in the ear ly grades.
1 0
Of course, effects also
could be offsett ing, or some combinations of placements could be more
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 99/375
8
FIGURE 5 1 First Grade Tracking Patterns: Repeaters
and Promoted Children Compared
First Grade Repeaters
Spec Ed +
Mid Read Gp
1.1
\
(N=88)
Promoted First Graders
N=508)
Spec Ed +
Low Read Gp
2.6
Low Read Gp
Spec Ed +
High Read Gp
2.2
Spec Ed +
Mid Read Gp
3.9
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 100/375
S
FIGURE 5.2 Low and High First Grad e Tracking Patterns
Firs t Graders with Low Placements
N=194)
Spec Ed + Low—
Read Gp
Ret + Spec Ed 3.0
Firs t G raders w i th High Placements
N=595)
Not Ret
High Read Gp-
-Not Spec Ed
-Not Ret
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 101/375
54
problem at ic than others . To add ress such issues , research w ill ha ve to beg in
examin ing the tracking system in its totality.
Th e lowe r cha rt of Fig ure 5.2 sh ow s the oth er side of first gr ad e tra ck ing .
O ver 90% of the coho rt (595/630) regis tered at least one no t lo w place
m en t. M ore than a third of this gro up , it tur ns out, w as favorably s i tuated
w ith respect to all three tracking cr i ter ia: these yo un gste rs we re in hi gh
reading groups, were not held back and did not receive Special Education
services . An other 40+%, wh i le not in high reading gr ou ps , nev er theless
m an ag ed to avoid both re tent ion and Specia l Educat ion. Hen ce , even in
Bal t imore publ ic schools , which enrol l mainly poor and minor i ty young
s te rs ,
most chi ldren are
not
tracked low ear ly on. The se yo un gs ters al l
moved on to second grade to realize whatever benefits follow a
high
track
his tory -promoted on schedule , top reading group and no ident i f ica t ion
with Special Education.
Placements Beyond First Grade
Table 5.2 exam ines the overlap in f irst gra de plac em ents in mo re detail .
It a lso sh ow s the pa ttern ing of later retention s and en cou nters w ith Special
Ed uca tion . Using data from school reco rds, retentio n and Special Edu ca
tion services are monitored through sixth grade, the f irs t year of middle
scho ol ; how ever because of re tent ion and other se tbacks , not ev eryo ne
m ak es i t to s ixth grad e on the sam e t imetable. The data in Table 5.2 take in to
account when tha t t rans i t ion was made, and so over lap the t rans i t ion to
m idd le school for everyo ne. Nine youn gs ters , or un de r
2% ,
skipp ed a grad e
in elem en tary school and w ere s ixth gra de rs af ter f ive yea rs;
6 1 %
(N=352)
w ere o n-time sixth gra de rs; 32.3% (N=185) w ere in sixth grad e in th eir
seventh school year ; and 4.7% (N=27) took eight years to make i t to s ixth
grade .
Table 5.2 reveals large, patterned differences associated with all three
dim en sio ns of f irst grad e tracking, long -term as well as sho rt- term . Th e left
s ide pa ne l covers ch ildren 's retention his tor ies . Alm ost 3 /4 th s of the
chi ldren in low fi rs t grade read ing gro up s w ere held back a t some point ,
over half in f irst gra de . A ddit ion ally, thr ou gh sixth gra de 35% w ere
re ta ined
a second time
In compar ison, none of the high reading group
you ng s ters repeated fir st grade , 88% ha d sm ooth prom otio n his tor ies and
jus t 6 (2 .5%) w ere dou ble repeaters by m idd le school . The interm edia te
level read ing g ro up yo un gs ters ' overa l l re tent ion ra te and do uble re tent ion
ra te both w ere betw een the high grou p- low gro up extremes .
1 2
T he m a pp in g
onto retention of children's relat ive reading group rank in f irs t grade thus
no t only is order ly ,
b ut extends all through the primary grad es
Specia l Educat ion a nd re tent ion a lso are l inked thro ug ho ut the pr im ary
gra de s. Special Ed uca tion s tud ent s , m ore likely to repeat firs t gra de , also
had a som ew hat higher re tent ion ra te in grades 2 -6 than did chi ldren wh o
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 102/375
TABLE 5.2 Ove r lap
Never Ret
1 . Fi rst Grad e Reading Gro up
Lowest 27.3%
(36)
Middle 56 .1%
(128)
Highest 87.7%
(207)
2. Special Education Year 1 or
Yes 27.6%
(29)
No 67.6%
(463)
3. Reta ined 1s t Grad e
Yes
—
N o 74.3
(492)
First Retention
Ret 1st
Gr a de
51.5%
(68)
8.8%
(20)
—
2
44.8%
(47)
11.8%
(81)
100.0%
(128)
—
*Note: The Years (rows) ver sus Grad es'
in Early Track Placements:
Ret Grs
2-6
21.2
(28)
35.1%
(80)
12.3%
(29)
27.6%
(29)
20.6%
(141)
—
25.7%
(170)
(columns) d
children according to their placements in Project Years 1 ar
Any 2nd Ret
34.9%
(46)
11.0%
(25)
2.5%
(6)
31.5%
(33)
9.8%
(67)
44.5%
(57)
6.5%
(43)
Fi rs t Grade
Never In
43.9%
(58)
79.4%
(181)
91.9%
(217)
—
87.3%
(598)
35.9%
(46)
83.4%
(552)
Through Sixth Grade
Special
Assign in
G rs
or 2
40.9%
(54)
11.8%
(27)
5.9%
(14)
100.0%
(105)
4.1%
(28)
46.9%
(60)
11.0%
(73)
Education*
Assign in
Grs 3-6
15.2%
(20)
8.8%
(20)
2 .1%
(5)
—
8.6%
(59)
17.2%
(22)
5.6%
(37)
Receive Ser
vices in Gr 6
50.9%
(43)
15.7%
(27)
6.0%
(10)
60.9%
(42)
13.5%
(64)
56.3%
(49)
12.5%
(57)
istinction is mea ningfu l for Special Educatio n. The first classifies
id 2. The second classifies children according to their
first , second, third, etc. More children are identified as receiving Special
2 because for the repeaters the Grade 1 or 2 per iod spans
grade level -
Education services in Grad es or 2 tha n in Years 1 or
as many as four years .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 103/375
8
w ere no t in Special Edu cation at the begin ning
1 3
and w ere m uc h m ore l ikely
to be held bac k a second t ime (31.5 ver sus
9.8 ).
First gra de rep ea ters also
w ere a t greater risk:
44.5
of them w ere held back a second t ime com pare d
to just 6.5 do ub le retention am on g child ren pr om ote d at the en d of firs t
grade (a fourth of whom were held back after f irst grade).
The r ight-s ide p ane l of Table 5.2 sho w s similar patter nin g w ith resp ect to
Special Education services beyond f irs t grade. Youngsters in low reading
groups and f irs t grade repeaters who made i t out of f irs t grade without
being assigned to Special Education s t i l l were more l ikely l ter to receive
services than we re children no t tracked low init ial ly. M ore than half those
w h o ha d low plac em ents in f irst grade received Special Ed uca tion services
in sixth gr ad e and th is ho ld s for all thre e varieties of first gra de tra ck ing . In
com par i son , the
high st
figure for chi ldre n not track ed low in first gra de w as
15.7 , for those in middle reading groups.
Sum mary: Early Tracking in First Grade and Beyond
To this point we have looked in detail at three dimensions of ear ly
t racking: reading group placement , re tent ion and ass ignment to Specia l
Edu cat ion. M ost you ngs te rs in the Beginning School S tud y were spare d
low p lace m ent s in al l three areas of tracking , b ut in f irst gra de 3 0 w ere low
in at least one area. M oreov er , for m an y childre n two or m or e of these
p laceme nts over lapped , and the comp oun ding of low p lacements ex tended
be yo nd f i rs t
grade .
Chi ldre n placed low in one area in firs t grad e we re m or e
likely also to occu py oth er low slots in f irst gra de , bu t even w h en m ult iple
low placements were avoided init ial ly the r isk of being placed low later ,
thro ug h s ixth grad e , wa s e levated. Educat ional t racking thu s begin s ear ly ,
involves severa l d imen s ions of t racking, touche s a grea t ma ny chi ldren in
var ious com binat ions and for m any ch i ldren has repercuss ion s a ll thro ug h
the pr im ary grades . The ques t ion addres sed next i s w heth er those reper
cuss ions extend to cur r iculum tracking in m idd le school.
Tracking in the M idd le rades
Curriculum differentiat ion in the middle grades s ignals the onset of
t racking as conven t ional ly un ders to od . At tha t point , s tud ents begin taking
different subjects (e.g., English versus reading) and the same subjects are
tau gh t at different lev els, often w ith little ov erla p in co nte nt. In m at h,
advanced students may take pre-algebra, while others s t i l l are working on
nu m be r skil ls ; in the lan gu age curr ic ulu m , high level s tu de nt s often take a
foreign language and may move on to l i terary cr i t icism and creative or
expository writ ing, while children in lower level classes are l imited to
remed ia l reading , rules of gramm ar , vocabu lary bui lding an d the l ike .
Un til m id dl e school, child ren 's op tions in pr incip le rem ain op en even if
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 104/375
8
they ha ve fallen beh ind . In high school, tho ug h, man y up pe r level cou rses
have prerequis i tes , and cur r icular decis ions in the middle grades are the
wa ysta t ion s to them . Ch i ldren placed in genera l m ath in m idd le school or
at the s tar t of high scho ol will not ta ke Calculus in twelf th gra de , no r w ill
chi ldre n taking funct ional re ad ing be able to enrol l in Ad van ced Placement
English li terature. Th e curr ic ulu m in the upp er grad es levels in the ma in
academ ic subjec ts i s sequen ced and hierarchical , so unless the gro un dw ork
ha s be en laid in m id d le scho ol stu de nt s are effectively cut off from hi gh lev el
op tion s (e.g. , G am ora n, 1992; O ake s, 1988; Oa kes, 19 89/90 ; Steven son,
Schiller an d Sc hneid er , 1994). We susp ect that tracking constrain ts also
extend do w n into the ear ly pr im ary grad es w here , we hav e seen, h id
d en forms of tracking touch m an y child ren's l ives . In this section, w e
examine how f irs t grade track placements ar t iculate with middle school
course placements in the language ar ts program ( reading and Engl ish) ,
math and fore ign languages .
Program Placements in Sixth Grade
BCPS schoo ls m ain ly follow a k-5, 6-8, 9-12 gr ad e org an iza tio n, w ith
m id dl e school beg inn ing in s ixth grad e. English displaces readin g as the
regu lar lang uag e ar ts cur r iculum in Bal t imore ' s m idd le schools , bu t in
s ixth gra de mo st BCPS s tude nts (about 3/4 ths) take both rea ding an d
Engl ish , a com m on p at tern na t ionw ide (Eps te in and M aclver , 1990). Al l
read ing courses are remed ia l an d /o r Specia l Educat ion, whi le the Engl ish
prog ram d is t ingu ishes am ong r emedia l , r egu la r Enr iched and Ad
vanc ed A cad em ic courses . T he rem edia l Engl ish cur r icu lum (e.g. , En
gl ish wi th Reading/Wri t ing Emphas is ) i s in tended for s tudents who fa i l
the Cit y 's readin g and w rit ing prof iciency tests. I t con centrates on p ara
graph construction, rules of grammar and the l ike, while Enriched and
A dva nced Ac adem ic s tud ents s tud y character iza tion, l i te rary devices and
au th or s ' pu rp os es , . .. w rit ing in a var iety of gen res, and som e low er level
skills of de ba te an d defens e (quoted from BCPS cou rse descr iption s) .
Enriched classes are available in al l m idd le schools; A dv anc ed Ac ade mic in
11 of 27, inc lud ing tw o city-wide m ag ne t schools .
The math program also dis t inguishes remedia l , regular and advanced
cou rses, the lat ter inv olv ing
pre-a lgebra /a lgebra sequence . These dis t inc
t ions are use d in m os t bu t not al l m idd le schools . For exa mp le, in on e school
atte nd ed by som e of the BSS yo un gste rs the re are no s ixth grad e classes as
such . Instead , s tu de nt s are placed in m ath an d English bas ed on their
assessed readin ess . C ours e levels in this school (5 in English an d m ath )
are distinguished according to level of difficulty and differ in student
com posit io n. A sixth gra der in a high er level cou rse that enrolls m ainly
seven th and e ighth grade rs w ou ld be cons idered adv anc ed, wh i le Level
4 an d 5 cou rses are low rega rdless of s tu de nt s ' gra de level . Even in this
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 105/375
FIGURE 5.3 Sixth Grade Course Placements by English Level
100
80
6 0 -
40
20
English Level
0 Rem edial (40.7 )
FJ Regular (38.8 )
0 Advanced (20.5 )
T
Lo Math
4
- •
J
1
Hi Math
i
Readingor. Lang.
school , then, i t i s poss ible to app roxim ate the rem edi a l - reg ula r
ad va nc ed dis t inc tions used genera l ly thro ug ho ut the Ci ty sys tem .
Foreig n la ng uag e instruction is available in all m idd le schoo ls , bu t not al l
s tu de nt s beg in a foreign lan gu ag e in s ixth grad e. In som e scho ols , for
ins tance , s tudents taking reading or remedia l Engl ish are not a l lowed to
take a foreign lan gu age . Since foreign lang ua ge s tud y in the m id dl e gra de s
has been found to inf luence program placements in high school (e .g. ,
A lexa nd er an d C ook, 1982; Ro sen bau m , 1976), thes e restr ict ions co uld be
con seq uen tial later . The re is no co un terp ar t of rem ed ial in the foreign
lang uag e pro gra m , but in som e schools ho no rs courses are dis t ing uish ed
f rom regu lar ones . Ho we ver , not m an y BSS you ngs te rs take high level
foreign language courses, so the comparisons that follow simply dis t in
guish an y fore ign lang uag e f rom no ne .
Table 5.3 descr ibes s ixth grade course level placements in English and
m ath ( low; regu lar ; high ) , bo th from the last qu ar te r available on tran
scripts,
14
and w he the r or not the s tu de nt took read ing or a foreign lan gu ag e
at any po int in the year . A lm ost 3/ 4t hs of the coho rt took reme dia l rea din g,
4 1 %
w ere in low level English,
20%
in low level m ath and a lmost 2 / 3r d s did
not take a fore ign lan guag e. Low level p lacem ents thus are com m onp lace ,
an d i t pro ba bly w ill com e as no su rpr ise that they ten d to al ign across areas
of the cur r iculu m . As disp laye d in Figure 5.3, for exa m ple, s tu de nts in
Enriched and Advanced Academic English are the ones most of ten ex
em pted f rom rem edia l reading : jus t un de r 39% of the 98 you ng s ters in
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 106/375
B
upper level English courses took reading ,
as
against
84%
of
the 94
students
in low level English and 92.4% of those in regu lar English. These youngs ters
also were m ore likely tha n the others to begin a foreign language in sixth
grade
(81%
compared to 9% of those in low level English classes and
41%
of
those in regular classes) and to be taking high level math
(52%
versus under
2%
of regular and low level English studen ts).
15
Figure 5.4 shows how these placements overlap or combine across all
four areas of the curricu lum , much as w as done in Figure 5.2 for the three
facets of first grade tracking. Although the tracking distinctions used here
are completely different than those in first grade, the idea at both poin ts is
to characterize students' relative standing in the then relevant hierarchies
of organ izational differentiation. Comparing just the sample splits in
Figures 5.2 and 5.4 we see many fewer high placements in sixth grade than
in first (90+% in first g rade
vs . 220/477,
or
46%
in sixth grade) and more low
placemen ts (30% vs. 88%).
w
It thus appears students' placements have
deterio rated from one level of schooling to the
next
at least in the aggregate,
although it remains to be determined whe ther the predom inant forms of
tracking at the two levels are at all comparable in terms of psychological
salience and practical import.
17
These are impo rtant issues, with potentially far reaching ramifications.
For example, children's school engagement, as reflected in their school
performance, liking of school, and academic self-image, all typically spiral
downward over the years. The reasons for this distancing of youth from
things academic are not well understood, but some research implicates
challenges encountered in acclimating to organizational arrangements in
the middle
grades,
including the system of tracking (for overview see Eccles
and M idgley, 1990; Eccles, Midgley and Adler, 1984).
Here we see that m ore children experience more low placem ents in sixth
grade than in first, and this alone could set them back if it shakes their
confidence and assaults their sense of self. But differences in the ch r cter
of tr cking
could further com pound ch ildren's difficulties. As typically
implemented, curriculum differentiation in the middle grades is akin to
whole class ability grou ping. Broad
scope,
exclusive and highly
visible,
this
is one of the m ost divisive kind of grouping arrangements
(e.g.,
Slavin, 1987;
Sorensen, 1970).
From the upper chart in Figure
5.4
we see that almost
15%
of the BSS sixth
grade rs with low placements were low across the boa rd. Other young
sters had just one (21.7%), two (27.0%) or three (26.8%) low placements,
18
with various language arts configurations dom inating the patte rns. The
isolated low placement most often is reading (21.7%); another frequent
pattern had reading trading off with foreign language study (27.0%), and
another com bined reading with rem edial English at the expense of foreign
language study
(22%).
Low math placem ents, in contrast, were both less
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 107/375
9
FIGURE 5 4 Low and High Sixth Grade Tracking Patterns
Sixth Grade Low Placement Patterns
N=419)
Other Combinations
Low Eng Low Math
No For Lang
4.8
Includes all patterns with
2.5
or
less
of
the sample
Sixth Grade High Placement Patterns
N=220)
High Eng Alone
For Lang
No Read
Other.
Combinations*
High Eng
For Lang
* Includes all patterns with
5
or
less
of
the sample
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 108/375
9
comm on and less concentrated.
High placements altogether were relatively rare, as more than half the
sample had
no
high sixth grade placem ents. At the same time, almost a fifth
of the cohort was placed high in two or more areas, so some youngs ters still
m anag ed to win d up with up pe r level concentrations (or, equivalently, to
avoid low level
ones,
e.g., reading). For
example,
93% of the students in high
level m ath also were in high level English,
90%
took a foreign language and
just
12.3%
took reading.
19
Altogether, almost
10%
of the cohort was high in
all four
areas,
which in Figure 5.4 we see accounting for a fifth of those with
any high placements.
That child ren's course levels tend to align across areas of the cu rriculum
is not surp rising. For one thing, children with strong academic records in
one area tend to be strong in other areas as well. Adm inistrative consider
ations also play a role. Most
schools,
for exam ple, use just a few tem plates
in scheduling classes. Block scheduling, in which pup ils move as a grou p
from class to
class,
simplifies plann ing bu t has the effect of placing children
in the same track for all or most of their academic subjects. But adm inis tra
tive actions can also force crossing
levels.
In Baltimore there are no remedial
m ath courses keyed to proficiency tests akin to English with R ea din g/
W riting Emp hasis in the language arts area. Instead, remedial courses in
math all are Special Education, so marginal studen ts are more likely to wind
up in low level English than in low level m ath.
Structural onstraints from First
rade
into Middle School
We now consider the overlap of sixth grade course-taking patterns with
first grade track placem ents. This is don e in Table 5.4, area by area and in
terms of low placements. Though individua l course-taking configurations
are too sparsely pop ulated to sustain com parisons, by tallying low and high
sixth grade placements we at least can determine whether constraints
associated w ith early tracking compound as the num ber of low first g rade
placements increases.
Associations betw een first grade tracking and sixth grade placements are
significant for all areas of the midd le grades curriculum save reading.
20
The
largest
Etas,
21
in the .3 - .4
range,
mainly involve placements in m ath and the
tally of low placements across all four areas, but the tie between reading
grou p level in first grade and English level in sixth grad e also is in this range.
While these relationships are only modest to m id-range, the enrollment
pa tterns tha t underlay them often are strikingly different. For example,
two-thirds of first g rade repeaters w ere in low level English in sixth g rade
compared to just over a third of prom oted first grad ers, a difference of 32.3
points revo lving around first grade retention. Larger still is the 35.2%
absolute difference
(61.5%
- 26.3 ) com paring children in high and low first
grade reading group s.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 109/375
TABLE 5.3 Sixth Gra de Course Placem ents : Engl ish Level , Math Level , Reading Enrol lment ,
Fore ign Langu age Enrol lment , and Tal l ies of Low and High Level Place ments
English
Low 40.7%
Level (194)
Regular 38.8
(185)
High
20 5
Level (98)
Ma t h
Low 20.2%
Level (97)
Regular 68.0
(327)
High 11.9
Level (57)
Reading
Taking 74.0%
(375)
Not 26.0
Taking (132)
Fore ign
La ngua ge
Taking 33.9%
(172)
No t 66 1
Taking (335)
# Low Level
Courses
0 12.2%
(58)
1 23.3
(111)
2
26 8
(128)
3 24.7
(118)
4 13.0
(62)
Includes low level English, low level math, taking reading, and not taking a foreign language
b
Includes high level English, high level math, not taking reading, and taking a foreign language
# High Level
Courses '
0 53.9%
(257)
1 24.7
(118)
2 7.8
(37)
3 4.4
(21)
4 9.2
(44)
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 110/375
9 1
At the high end, advanced s ixth grade placements were rare among
youngsters tracked low in f irs t grade, especially f irs t grade repeaters , just
tw o of w ho m w ere in high level s ixth grade English, one in hig h level mat h,
By w ay of com par iso n, of the young s ters prom ote d the fir st t ime thro ug h
first gra de , a fou rth as s ixth gra de rs we re in hi gh level Eng lish an d a s ev en th
w ere in hig h level m ath . Ch ildre n in low read ing gr ou ps in f irst gra de or in
Special Ed uca tion w ere no t exclud ed from high level cou rses in s ixth g rad e
to the same extent as f irs t grade repeaters , but such placements s t i l l were
relatively rare for them . For exa m ple, the absolute difference co m par ing
high and low f irs t grade reading groups in terms of s ixth grade high level
English en rollm en ts (38.0 -12.8 =25.2 ) s close to tha tbe tw een re ta ined
and pro m ote d yo un gs ters (22.4 ).
Table 5.4 th us sh ow s the relevance of chi ldren 's tracking exp erience s in
f irs t grade to s ixth grade placements area by area. The r ightmost columns
show addi t ional ly tha t p lacements cross the four areas also map onto first
gra de track ing. Chil dre n in Special Educa tion in f irst g rad e, for instance, on
ave rag e took 0.9 mo re low level cou rses and 0.5 fewer hig h level cou rses as
sixth graders than did their cohortmates not in Special Education, and a
sim ilar patt ern h ol d s for the other facets of first grad e trackin g. Since first
grade track placements overlap substantial ly, this does not necessar i ly
m ea n that al l three dim en sio ns of ear ly tracking con tr ibute un iqu ely to s ixth
gra de cou rse taking pat tern s , bu t a t the zero-order level connect ions involv
ing a l l three—retent ion, Specia l Educat ion and reading group level—are
significant, w ith m os t in the .20 - .35 range .
22
Finally, the last entr ie s (bottom pan el) in Table 5.4 sh ow ho w sixth gr ad e
curr icu lar place m en ts m ap o nto the tal ly of low f irst grad e place m ents . If,
as m igh t be suspected , the consequ ences of ear ly t racking co m po un d across
types of tracking, then the l ikelihood of being placed low in s ixth grade
sho uld increase as the nu m be r of low firs t gra de place m ents increases. W ith
the exception of rea din g, this seem s to be the case. Ind eed , yo un gst ers low
on all thre e ear ly trackin g mea sure s consis tently fared wo rst in s ixth grade.
23
Ho wev er , those wi th tw o low f irs t grade placem ents , comp ared to the low
groups on each of the
individual tracking measures
genera l ly do not have
either the highest percentage of low sixth grade placements or the lowest
perc enta ge of high s ixth grad e placem ents . In m ost instances there is at least
one other group—sometimes repeaters , somet imes members of low f i r s t
gra de readin g grou ps , som et imes chi ldren ass igned to Specia l Educat ion—
with more extreme percentages .
These are small dif ferences, bu t they s igna l som e of the com plexit ies tha t
ar ise as one begin s to look broa dly at system s of tracking . Tall ies of low an d
high placements , as used here , useful ly summar ize chi ldren 's genera l
s tan din g as it ove r laps indiv idual t racking hierarchies, bu t it app ears som e
placements matter more than others , at least for some areas of the middle
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 111/375
First Grade
Placements
Special
Education
Yes
N o
Eta
Retained
Yes
N o
Eta
% Low
62.5
37.3
67.0
34.7
TABLE 5 4
English
% Reg
29.7
40.2
.18***
30.7
40.6
.28***
Level
% Hi gh
7.8
22.5
2.3
24.7
Ove r lap of Firs t Gra de and Sixth Grad e Track Placemen ts
Sixth Grade Course Levels and Course taking Pat terns
(N)
(64)
(413)
(88)
(389)
% Low
53.1
15 1
50.6
13.3
Math Level
% Reg
42.2
71.9
.28***
48.3
72.4
.35***
% High
4.7
12.9
1.1
14.3
(N)
(64)
(417)
(89)
(392)
a
a
Taking
For Lang
(N)
15.7
(70)
36.8
(437)
.15***
9.6
(94)
39.5
(413)
.25***
A
Ta k i ng
Reading
(N)
72.9
(70)
74.1
(437)
. 0 1
N S
81.9
(94)
72.2
(413)
.09
NS
Low
Cour s e s
(N)
2.8
(64)
1.9
(413)
.24***
2.9
(88)
1.8
(389)
.36***
4 i
High
Courses
(N)
0.5
(64)
1.0
(413)
.13**
0.3
(88)
1.0
(389)
.23***
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 112/375
Reading Group
Lowest 61.5
nt r-
mediate
Highest
Eta
# Low
Placements
0
1
2
3
Eta
Significance
43.1
26.3
32.3
41.0
70.0
77.8
• levels:
25.6
46.4
35.8
. 3 3 *
42.6
41.0
22.5
22.2
.28 *
12.8
10.5
38.0
25.1
18.0
7.5
0.0
(78)
(153)
(137)
(263)
(61)
(40)
(18)
*p < .05 level; **p < .01 level;
50.0
16.7
7.3
9.4
27.9
50.0
77.8
p<
43.6
77.6
70.8
.38***
75.9
63.9
45.0
22.2
.41***
.001 level
6.4
5.8
21.9
14.7
8.2
5.0
0.0
(78)
(156)
(137)
(266)
(61)
(40)
(18)
19.3
(83)
25.1
(167)
44.8
(145)
.23***
39.1
(279)
29.0
(69)
14.0
(43)
0.0
(18)
2^***
73.5
(83)
79.0
(167)
65.5
(145)
4
N S
73.5
(279)
68.1
(69)
76.7
(43)
83.3
(18)
.07
NS
2.7
(78)
2.2
(153)
1.6
(137)
.35***
1.8
(263)
2.1
(61)
2.9
(40)
3.4
(18)
.37***
0.6
(78)
0.6
(153)
1.4
(137)
.30***
1.0
(263)
0.8
(61)
0.5
(40)
0.2
(18)
.19**
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 113/375
9
grad es cur r iculum . W e a lso suspect tha t p lacemen t conf igura t ions wi l l
have different consequences depending on exactly how they f i t together
organiza t ional ly an d adminis t ra t ive ly . In other w ord s , research need s t be
sensit ive to no nad dit iv it ies or con dit iona l relat ionships w he n trying to sor t
ou t effects invo lving multip le track hierarchies . These com plication s de
serve fuller treatm ent than is possible her e.
Discuss ion
Educational tracking exists in many forms at al l levels of schooling.
Resea rch on curr i culu m differentiation in hig h school init ial ly w as framed
in terms of over -arching programs: col lege prepara tory, vocat ional , gen
eral , an d such. Such global dis t inctions rema in useful for som e pu rp os es ,
bu t co urse takin g pa tter ns increasingly cross these tradit ion al lines of div ide
(DeLany , 1991; G aret an d D eLany, 1988; Pow el l , Far rar and Co hen , 1985)
and i t has proven useful in recent research to examine instead placement
practices and consequences in specif ic areas of the curr iculum, l ike the
lan gu ag e ar ts , m ath , or science. Doing so reveals that de term ina nts of hig h
level and low level course placements sometimes differ across subjects
(G am oran , 1992; Ha ll inan, 1992; Stevenson , Schil ler and Schn eider , 1994) ,
that co nse qu enc es of track level also differ (Hoffer, 1992), an d th at track ing
ar ran gem ents themselves are sensi t ive to school organiz a t ional p rope r t ies ,
instructional practices and classroom processes (Bryk and Thum, 1989;
Kilgore, 199 1; Kilgore and Pen dleto n, 199 3; Lee and Bryk, 1988; Lee an d
Smith, 1993).
Progress on these many f ronts is leading to a more mature , nuanced
un de rs ta nd in g of ho w t racking ac tually wo rks . H ow ever , the disag gre
gated approach car r ies l imi ts of i t s own and no doubt s tudies soon wil l
begin reag greg at ing the cur r iculum , a l tho ug h not in the w ay it used to be
done .
2 4
Look ing at course plac em ents pieceme al, as curre nt practice en
co ura ge s, neglects t ies across areas of the cur r icu lum tha t give tracking i ts
sys temic character. The t radi t ional pro gra m dis t inc t ions lum pe d too m uch
toge ther , but a cou rse level focus loses s ight of pr og ram coo rdin ation , an d
so in i ts own way also obscures important aspects of tracking.
These sam e concerns app ly to ear ly tracking. De spite the preten se of a
uniform curr iculum, children's experience of schooling in the pr imary
gra de s is far f rom un iform. The re are any nu m be r of org ani zatio nal an d
adm inis t ra t ive prac t ices a t the school and c lass room level tha t use s up po s
edly educat ional ly re levant cr i te r ia to dis t inguish chi ldren f rom one an
other . These create new organizational identi t ies , s tructure the daily rou
t ine,
and build in differences where otherwise none might exis t .
Abi l ity grou pin g in the pr im ary g rades mo st clear ly para l le ls c ur r iculum
differentiation at the secon dary schoo l as a m o d e of org aniz ation al differ-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 114/375
9
entiation
(e.g.,
Rosenbaum,
1980; 1984),
yet just as there are many different
tracks in the up pe r grades, so too is early tracking multifaceted. Reading
groups, retention and Special Education are three prototypical forms.
Achievement outcomes in the upper grades (e.g., test scores; dropout)
can be predicted from early schooling markers (Cairns, Cairns and
Neckerman, 1989; Ensminger and Slusarcick, 1992; Lloyd, 1978), so how
child ren settle initially into the student role is of great consequence over the
long hau l. Elsewhere we have argued the importance of the beginning
school transition as a critical period for children 's schooling (Entwisle and
Alexander, 1989; 1993). Some of the reasons for the crucible na ture of early
schooling involve considerations internal to the child (e.g., the rap id pace
and chang ing natu re of cognitive and affective developm ent in the 6 -8 age
range).
These reasons are widely recognized, but other reasons, less well
appreciated, involve considerations internal to the school, including the
largely buried systems of tracking that begin to channel children along
different educational pathways right from the start.
This paper has described first grade tracking and how it ties in with
course taking in sixth
grade.
Desp ite extensive literatures on all three of the
tracks examined here, their overlap has been virtually ignored and this
could be a serious oversight. About
30%
of the
SS
cohort were placed low
on at least one of the three
tracks;
14%
w ere tracked low in two or m ore
areas.
Furtherm ore, low placements combined in just about all ways po ssible, and
children placed low in first g rade w ere at elevated risk later of being held
back either for the first or second time or of being assigned to Special
Education.
Many youngsters thus occupy low slots in the school's tracking hierar
chies at the very start of their school careers. Though o ur research is
Baltimore based, virtually all schools gro up in read ing or other subjects so
the problem is not limited to poor, urb an a reas. All these placements are
intended
to
help ch ildren, and
it
should not be assum ed their consequences
are necessarily negative. Our inquires and those of others, for exam ple, find
slower than expected academic progress among children placed in low
reading groups in first grade (e.g., Gamoran, 1986; Pallas, Entwisle, Alexander
and Stluka,
1994),
bu t for children in the
SS
retention has mainly positive,
not negative, effects (Alexander, Entwisle and Dauber, 1994).
25
Moreover,
and of most immediate relevance, there are important issues involving
early tracking that have received virtually no attention.
For one, whether particular tracking configurations have distinctive
consequences is unkn ow n, but our overview of first grade profiles identifies
multiple tracks as a practical concern. Retention, Special Education and
reading group standing are not isolated events in children's experience.
They intersect in all sorts of ways . Though low stan ding in one area
increases the odd s of low standin g in the
others,
placem ents across the three
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 115/375
9
dimensions of f irs t grade tracking are far from perfect ly al igned: some
repea ters a re not in low read ing gro ups or ass igned to Specia l Edu cat ion;
despi te chi ldren ' s be ing in low reading groups , many are promoted a t
ye ar 's en d an d d o no t receive Special Ed ucatio n services; ther e even are
Special Ed ucation s tu de nts with no other low track identi t ies . For m an y
chi ldre n t racking me ssages thus a re mixed , and th is am bigui ty c ould soften
the impact of negative messages that revolve around their low track
identif ication. O n the other ha nd , m ult iple low track pla cem ents co uld
pr ov e to be an especial ly w eighty bu rde n.
A second op en quest io n is h ow e arly trackin g t ies in w ith track ing
sys tem s a t the up pe r grade leve ls . T rack p lacem ents cons t ra in o pp ort un i-
t ies a t the m idd le gr ad es and be yo nd (e .g. , G am ora n, 1992; H all ina n, 1992),
bu t w he the r such cons t ra in ts extend back to chi ldren ' s very f irst enco unter s
w i th educat iona l t racking rem ains to be de term ined . M ost eva lua t ions of
early tracking examine consequences for school performance or in the
s o c i o e m o t io n a l r e al m ; h o w e v e r , t ra c ki n g s y s t e m s d o n o t j u s t c h a n g e y o u n g -
sters ,
they also channel them, opening doors for some, closing them for
others (e.g., Oakes, 1988; 1989/90).
Structural constraints originating in tracking systems at the earl ies t
grade levels are just beginning to command at tention. Kerckhoff 's recent
s tudy of within and between school tracking in Great Bri ta in (1993), for
exa m ple , d iscovered subs tant ia l ins t i tu t iona l iner t ia in m ath p lacem ents
from early elementary school (age 7) to late elementary school (age 11).
C hild ren in tracked m ath classes (about 30% of the total) w ere l ikely to be
placed s imilarly on both occasions,
26
and s t ruc tura l connect ions across
levels of schooling ev en w ere detect ible at age 16, the las t year of co m pu l-
sory schoo l at te nd an ce in Grea t Brita in at the t im e (1974). W he ther the s am e
sort of inst i tut io nal inert ia a lso characterizes early tracking in the U.S.,
an d w he the r facets of t racking other tha n abili ty gr ou p level are im plicated ,
rem ain to be de te rm ined .
To our knowledge the data presented in the present paper are the f irs t
documenta t ion of connect ions be tween ear ly t racking and middle school
curr iculum placements tha t encompass severa l h ie rarchies of organiza-
t ional differentiat ion at bo th levels of schooling . De script ive detai l w as
provided on the overlap of f irs t grade retention, reading group level and
Special Education with la ter retention, la ter receipt of Special Education
services , and s ixth grade (middle school) placements in English, reading,
m ath a nd fore ign languag es . O ver lap in p lacem ents from f irst to s ix th g rad e
w as observ ed for a l l a reas of the m idd le gra des curr icu lum save rea ding .
W hile cons iderable add i t iona l w ork wi l l be requ ired before we can ta lk
in term s of effects as dis t inct from pa tter ns , the orde rl iness seen here in
placements s tretching from the very s tart of e lementary school into the
m idd le grad es i s cons is tent wi th the v iew of tracking sys tem s as com pris ing
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 116/375
oppor tuni ty s t ruc tures tha t advance the in teres ts of some whi le impeding
tho se of oth ers. I t is strikin g tha t action s take n in first gra de cou ld h av e
conse quen ces so far rem ove d in t ime ( 6 - 8 years ) an d space ( in mi dd le
school as the ven ue ) , bu t the persis tence o ver t ime of hierarc hies of inequ al
ity is one of the defining features of systems of stratif ication, and educa
t ional inst i tutions are no more exempt f rom strat ifying processes than are
any of society 's oth er insti tution s. The stratif ication em bo die d in ear ly
tracking thus far has been largely hidden from view, however , and i ts
conseq uences s ti ll a re not wel l u nd ers to od .
Notes
Data collection for this research was supported by the W.T. Grant Founda tion
Grant N o. 83079682 and National Institute of Child Health and Deve lopmen t Grant
No.
1R0116302.
The
analysis
was
supported by National Science Foundation G rant
No. SES 8510535 and National Institute of Child Health and Development Grants
No. 1 R01 21044,5 R01 23738, and 5 R01 23943; and by Grant No. R117D4005 from
the Office of Educational Research and Improvem ent, U.S. Department of Educa
tion, to the Johns Hopkins University Center for Research on the Education of
Students Placed at Risk. We thank the children, parents, teachers, principals and
other school system personnel who have given u s such splendid cooperation in all
phases of this research.
1.
Just over
1
(N=7) were accelerated. According to more recent data (th rough
spring 95 or 13
years),
about half the cohort has grad uated high school, a fourth has
drop ped out and another fourth are still in school, pu tting them at least one year
behind and thus at elevated risk of drop out.
2.
For additional detail on these pathw ays thro ugh the first eight years of the
cohort's experience, see Alexander, Entwisle and Dauber,
1994,
Chapter 2.
3.
Considering just ability grou ping in the primary
grades,
one reviewer (Sla vin,
1987) distinguishes seven different grouping arrangements.
4.
Care also must be exercised no t to generalize conclusions across different type s
of grouping arrang em ents (e.g., Reum an, 1989; Rowan and Miracle, 1983).
5. SeeO akes, Gam oran and
Page, 1992,
for an
excellent
comprehensive overview
of these and other issues involving tracking at all levels of schooling.
6. See Karweit, 1992, for a balanced overview .
7.
Spring reading grou p
level is
know n for
596
m embers of the
cohort. These
data
come from teachers. In the case of Special Education, data from the first tw o years
are combined. W e initially sampled regular classrooms, so hardly any youngsters
were in separate classes the first year (just 2 of
790,
as com pared to 63, or 8 , in
pullout-program s). This clearly is an undercount; and since assignments in a given
year frequently are based on assessments made the previous spring, we reasoned
that including second year placements in the tally would
give a
fairer reading of the
early Special Education experience. The 13 figure includes children wh o received
supplem enta l services from regular classes (i.e., pullout programs in reading a n d /
or math) as well as those assigned to separate Special Education classes.
8. Thirty-four o the rs were in classes that did not use small groups for reading ,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 117/375
1
an d so cou ld not be classified in term s of ran k position. N o inform ation is ava ilable
for the other 160 m em bers of the cohort (20% m issing data). The se data com e from
50 first gra de teachers w ho res pon ded to our inq uiry about the gr ou p placem ents of
the BSS youngsters in their classes.
9 . This app roa ch sl ides ove r dist inct ions in the m idd le ranks, bu t ident ifies
chi ldren placed highest and lowest , where the signals that at tach to group place
m ents should be c learest . The most comm on arrangem ent wa s th ree gro up s , used
by 29 of the
5
t eachers . The next mo st com mo n pa t te rn w as four g roup s , em ployed
by 12 teachers. Four teachers used just two groups, while three used five groups.
This dist ribut ion seems pret ty typical of pract ices from the early eight ies (e .g . ,
Dr eebe n, 1984; Ha ll inan and S arensen , 1983). Sixty percent of the stu dy you ngs ters
w ere in classes w ith three grou ps ; a fourth w ere in classes w ith four gr ou ps . The
other a rran gem ents , inc lud ing ungro upe d c lasses , involved fewer than ten percen t
of the sample.
10. The real i ty of mult iple low p lacem ents also com plicates al locat ing resp onsi
bi l i ty when trying to assess how any one dimension of t racking affects chi ldren 's
schooling. Est im ates of ho w re ading gr ou p level affects ach ievem ent , for exam ple,
could be off unless effects associated with retention or Special Education are
adjusted for.
11. This inform ation is avai lable only for chi ldren w h o rem ained in the BCPS the
en tire t im e, 72.5% of the total (N =573).
12.
The sam e holds if the figures in Tab le5.2 are recast in terms of relat ive risk.
Ther e a re 64 low g ro up chi ldren w ho we re not held back in first grad e. The se are the
low gr ou p chi ldren at r isk of f irst retent ion after fi rst gra de. Tw enty-e ight of them ,
o r 43.8%, w ere he ld back in grade s 2 - 6 . This com pares to
38.5%
of the 208 m idd le
gro up ch i ld ren wh o were no t first g rad e repea ters .
13.
Ag ain, relat ive risk com pariso ns poin t in the sam e direct ion. Half the first
gr ad e Special Educ at ion ch i ldren no t held back in first g ra de are held back later (29 /
58) com pare d to 23.3% of those not in Special Education (141/604).
14 . Som e s tuden ts transferred ou t dur ing the year and som et imes courses w ere
not take n for the full yea r. Referencing enr ollm en ts to the last availa ble qu ar ter
m axim izes case coverage .
15.
Had we used math placements as the frame of reference, differences across
o ther a re as of the curr icu lum w ould have been even la rger.
16 . The sam e comp ariso ns also can be seen in the low an d high tal l ies re po rted
in the rightmost columns of Tables 5.1 and 5.3.
17 .
Also , s ince w e cover abi li ty gro up ing in just one area, the first gr ad e pa t ter ns
l ike ly und ers ta te low p lacem ents . We dou bt tha t mo re comp le te coverage wo uld
al ter the picture appreciably, however.
18 .
T he other com bina t ions total of just un de r 10% w ou ld ad d, at m ost , 2% to
the se totals. For exam ple , 2.0% of the 419 wer e not taking a foreign lang uag e as their
on ly low p lacement .
19 .
Math is m ore select ive than English w he n i t com es to high level placem ents,
so all thes e figures are lowe r usin g English as the frame of reference. Still , the tren d
is similar (see Figure 5.3).
20. Seventy-four percent of the sample takes reading in sixth grade, so the
d is t r ibu t ion skew no doubt damps re la t ionsh ips involv ing read ing .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 118/375
101
21.
Calculated from the cross-classifications that also are the source of the
percentage d istributions in Table 5.3, Eta is a measure of association analogous to the
product-moment correlation but applicable to nominal and ordinal as well as
interval level measurements.
22.
Indeed, in m ost instances the associations, though attenuated, hold up even
wh en controls are introduce d. Two of the six partial correlations invo lving Special
Education services (net of retention and reading g roup level) are significant at least
at the
.05
level
(
low placements, .12; math level, .16); fiveof six involving retention
status are significant ( low courses, .19; high courses, .12; English level,
.13;
math
level, .16; foreign language enrollment, .14); and five of six involving read ing grou p
level are significant (# low courses, .19; # high courses, .15; English level, .19; math
level,
.22;
foreign language enrollmen t, .10).
23.
This, though, is a small group (N=18 in Table 5.3) and their situation is
extreme. Beyond their standing as members of low reading g roup s, first grade
repeate rs and recipients of Special Education services, their academ ic profile in first
grade (test scores and marks) puts them far below all other groups. Und er such
extreme circumstances, sorting out the sources of their academic difficulties later on,
including the reasons behind their low midd le school placements, will be exceed
ingly difficult.
24. See Stevenson, Schiller and Schneider, 1994, for an example of the sort of
approach we have in mind.
25. Even where the w eight of evidence suggests generally adverse effects (i.e., for
children in low ability groups), exceptions to the rule often can be found (e.g.,
Gamoran, 1993).
26.
With sizeable, significant regression effects pre dictin g later group level from
earlier group level, net of a whole host of controls.
References
Abidin, R.R., W.M. Golladay, and A.L. H owerton. 1971. Elementary School Reten
tion: An Unjustifiable, Discriminatory and Noxious Policy.
Journal of School
Psychology
9:410-17.
Alexander, K.L. and M.A. Cook. 1982. Curricula and C oursew ork: A Surprise
Ending to a Familiar Story.
merican Sociological Review
47:626-40.
. D.R. Entwisle, and
S
Dauber. 1994. On the Success of Failure: A Reassessment
of
the Effects
of
Retention
in the
Primary
Grades Cambridge, MA: Cambridge
University Press.
Algozzine, B, and J. Ysseldyke. 1983. Learning Disabilities as a Subset of School
Failure: The Over-Sophistication of a Concept.
Exceptional Children
50:242-46.
Archambault Jr., F.X. 1989. Instructional Setting Features and O ther Design Fea
tures of Com pensatory Education Program s.
Pp.
220-63
in
Effective Program s for
Students at
Risk edited by R.E. Slavin, N.L. Karweit and N.A . Ma dden. Boston:
Allyn and Bacon.
Bianchi,
S
1984.
Ch ildren 's Progress Through School: A Research
Note. Sociology
of Education
57:184-92.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 119/375
2 2
Bryk, A.S. an d Y.M. Th um . 1989. The Effects of H igh School O rg an iza tion on
Dropping Out : An Explora tory Inves t iga t ion . American Educational Research
Journal
26:353-83.
Ca irns, R.B., B.D. C aim s, and H.J. N ecke rm an. 1989. Early School Dr opo ut:
Configura t ions and Determinants . Child Development 60:1437-52.
Ca rrie r, J.G. 1986. Sociology and Special Ed uca tion: Differentiation an d A llocation
in Mass Educat ion .
American Journal of ducation
94:281-312.
D eL any , B.D. 1991. Alloc ation, Choice, an d Stratification W ithin Hig h Schools:
How the Sort ing Machine Works .
American Journal o f Education
99:181-207.
Dree ben , R. 1984. Fi rs t -Grade Read ing Gr oups: Thei r Form at ion and C han ge . Pp .
69-84 in The Social Context of Instruction: Group Organization and Group Process
ed i ted by P.L. Peterson, L.C. W ilkinson and M . Hall inan . San Diego: A cade m ic
Press.
Eccles, J .S. an d C. Midg ley. 1990. Ch ang es in Ac adem ic Motivat ion an d
Self
Percept ion D urin g Early Ad olescen ce. Pp . 134-55 in From Childhood to Adoles-
cence: A Transitional Period? ed i ted by R. Mo ntema yor , G.R. Ad am s an d T.P.
Gu lo t ta . N ew bury Park , CA: Sage .
, C. Midg ley, and T. Ad ler. 1984. Gra de-R elated C han ges in the School
En viron m ent: Effects on Ac hievem ent M otivat ion. Pp. 283-331 in The Develop-
ment ofAchievement Motivation
edi ted by J.G. Nicholls . Gr eenw ich,
C T:
JAI Press .
Ens m inger, M.E. and A .L. Slusarcick. 1992. Path s to H igh School G rad ua t ion or
Dropout: A Longitudinal Study of a First Grade Cohort . Sociology of Education
65:95-113.
En tw isle, D.R. an d K.L. A lexa nde r. 1989. Early Schooling as a 'Critical Pe rio d'
Ph eno m eno n. Pp . 27-55 in
Sociology of Education and Socialization
edi ted by K.
N am bo od iri and R.G. Corw in, Greenw ich, CT: JAI Press.
and K.L. A lexan der. 1993. En try Into Schools: The Beg innin g School
Tran si t ion and E ducat iona l Strat if icat ion in the U nited States. Pp . 401-23 in
Annual Review of Sociology vol . 19. Palo Alto, CA: An nu al Review s, Inc.
Ep stein, J.L. and D.J. M aclver. 1990. Ed ucatio n in the M idd le G rad es:
National Trends
and Practices. Colu m bus, OH : Nat ional M iddle School Associa tion .
Fine, M.
1991.
Framing Dropouts: Notes on the Politics of an Urban Public High School.
Albany, NY: State Universi ty of New York Press.
G am or an , A. 1986. Instruct ional and Inst i tut ional Effects of Abil i ty G ro up in g.
Sociology of Education
59:185-98.
. 1992. Access to Excellence: A ssig nm ent to H on or s Eng lish Cla sses in the
Transi t ion to High School .
Educational Evaluation and Policy Analysis
3:185-204.
. 1993. Altern at ive Uses of Abil i ty Gro up ing in Seco ndary Schools: Ca n W e
Bring High-Quali ty Instruct ion to Low-Abil i ty Classes? American Journal of
Education
102:122.
Garet , M. an d B.D. DeL any. 1988. Stud ents, Co urse s and Strat i ficat ion.
Sociology
of Education 61:61-77.
Haller, E.J . 1985. Pupil Race and Elem entary School Abil i ty G rou pin g: Ar e Teach
ers Biased Aga inst Black Ch ildre n? American
ducational
Research Journal22:456-
83.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 120/375
103
and S.A. Davis. 1980. Does Socioeconomic Status Bias the Assignment of
Elementary
School
Students to Reading Groups? merican Educational Research
Journal 17:409-18.
Ha llinan, M.T. 1992. The Organization of Students for Instruction in the M iddle
School. Sociology of Education 65:114-27.
and A .B. S0rensen. 1983. The Formation and Stability of Instructional
Groups. merican Sociological Review 48:83851.
Hauser, R.M. 1970. Educational Stratification in the United States. Pp . 102-29 in
Social
Stratification:
Research and
Theory for
the
1970s
edited by E.O. Laumann.
New York: Bobbs-Merrill.
Heller, K.A., W.H. Ho ltzma n, and S. Messick, eds . 1982. Placing Children in Special
Education: A Strategy for
Equity. Washington,
DC:
National Academy Press.
Hoffer, T.B. 1992. M iddle School Ability Grouping and Student Achievement in
Science and Mathematics. Educational Evaluation and Policy nalysis 14:205-27.
House, E.R. 1989. Policy Implications of Retention Research. Pp. 202-13 in lunk-
ing Grades:
Research and Policies on
Retention
edited by L.A. Shepard and M.L.
Smith. London: Falmer Press.
Karw eit, N . 1992. Reten tion Policy. Pp . 1114-18 in Encyclopedia
of
Educational
Research edited by M. Alkin. New York: Macmillan.
Kelly, S.P.
1989.
15,000 City Students Failed Despite Debate Over P rom otions. The
Evening Sun 25 July, C I.
Kerckhoff A .C
1993.
Diverging Pathways: Social Structure and Career Deflections. New
York: Cambridge Press.
Kilgore,
S.B.
1991. The Organizational Context of Tracking in Schools. merican
Sociological Review 56:189-203.
and W.W. Pend leton. 1993. The Organ izational Context of Learning:
Framework for Understanding the Acquisition of Knowledge.
Sociology
of
Education 66:63-87.
Kulik, J. A. and C.L. Kulik. 1987. Effects of Ability Grouping on S tudent Achieve
ment. Equity and Excellence 23:22-30.
Lee, V.E. and A.S. Bryk. 1988. Cu rriculum Tracking as Mediating the Social
Distribution of High School Achievement. Sociology
o f
Education 61:78-94.
and J.B. Smith. 1993. Effects of School Restructuring on the Achievem ent
and Engagement of Middle-Grade Students. Sociology of Education 66:164-87.
Leinhardt, G. and A . Pallay. 1982. Restrictive E ducational Settings: Exile or Ha
ven?
Review
of
Educational Research 54:557-78.
Lloyd,
D.N.
1978. Prediction of School Failure from T hird-Grade Data. Educational
and Psychological Measurement
38:1911200.
M adden, N.A. and R.A. Slavin. 1983. M ainstreaming Students with Mild Handi
caps:
Academic and Social Ou tcom es. Review
of
Educational Research 53:519-69.
McPartland, J.M., J.R. Coldiron, and J.H. Braddock. 1987. School Structures
and
Classroom Practices in Elementary
Middle
and Secondary
Schools. Report No . 14. The
Johns Hopkins University, Center for Research on Elementary and Middle
Schools.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 121/375
2 4
Mehan, H. 1992. Unde rstand ing Inequality in Schools: The Contribution of Inter
pretive Studies.
Sociology
of
Education
65:1-20.
, A. Hetweck, and J.L. Meihls. 1985.
Handicapping the
Handicapped:
Decision
Making in Students Careers.
S tanford, CA: Stanford University P ress.
Mercer, J
1974. Labelling theMentally
Retarded Berkeley, CA: U niversity of California
Press.
Oakes, J. 1988. Tracking in Mathematics and Science Education: A Structural
Contribution to Unequal Schooling. Pp . 106-25 in
Class, Race and Gender in
merican
Education, edited by L. Weis. Albany, NY: State University of New York
Press.
. 1989/90. Op portunities, Achievement and Choice: W omen and M inority
Studen ts in Science and Mathematics. Review of Research in Education 16:153-222.
. 1992. Can Tracking Research Inform Practice? Technical, Norm ative and
Political Considerations.
Educational Researcher
21:12-21.
, A. Gamoran, and
R N
Page.
1992. Curriculum Differentiation: Opp ortu ni
ties,
Outcomes and M eanings. Pp.
570-608
in
Handbook of Research on Curriculum,
edited by
P W
Jackson. New York: Macmillan.
Office of Special Education Programs. 1991.
To Assure the Free ppropriate Public
Education of ll Children with Disabilities: Thirteenth nnual Report to Congress of the
Implementation
of the
Individuals
with
Disabilities
Act. Washington, DC: U.S.
Department of Education.
Pallas, A.M.,
D R
Entwisle, K.L. Alexander, and M.F. Stluka. 1994. Ability-Group
Effects: Instructiona l, Social or Institutional?
Sociology o f Education
67:27-46.
Peterson, S.E.,
J.S.
DeGracie, and C R. Ayabe. 1987. A Longitudinal Study of the
Effects of Retention/Promotion on Academic Achievement.
merican Educa-
tional Research Journal
27:107-18.
Pierson , L.H. and J.P. Connell. 1992. Effect of Grade Retention on Self-System
Processes, School Engagement and Academic Performance.
Journal of Educa-
tional Psychology
84:300-07.
Pow ell, A., E. Farrar, and D.K. Cohen. 1985.
The Shopping Mall High
School Boston:
Houghton-Mifflin.
Reum an, D.A. 1989. How Social Com parison M ediates the Relation Between
Ability-grouping Practices and Students' Achievement Expectancies in Math
ematics.
Journal o f Educational Psychology
81:178-89.
Reynolds, A.J. 1992. Grade Retention and School Adjustment: An Explanatory
Analysis.
Educational Evaluation and Policy nalysis
14:101-21.
Reynolds, M.C. 1984. Classification of Students with H and icaps. Review o f Re-
search
in
Education
11:63-92.
Rist, R. 1970. Social Class and Teacher Expectations: The Self-Fulfilling Prophecy
in Ghetto Education. Harvard Educational Review 40:411-51.
Rosenbaum, J.E. 1976. Making Inequality. New York: Wiley.
. 1980. Some Implications of Educational Grouping .
Review of Research in
Education
8:361-401.
.
1984. The Social
O rganization of Instructional G rouping. Pp.
53-68
in
Tire
Social Context of Instruction: Group Organization and Group Process, edited by P.L.
Peterson, L.C. Wilkinson and M. Hallinan. San Diego: Academ ic Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 122/375
105
Rowan, B. and A W Miracle. 1983. Systems of Ability Group ing and th e Stratifica
tion of Achievement in Elementary Schools.
Sociology
of
Education
56:133-44.
Shepard, L A an d
M.L. Smith. 1989.
Introduction and Overview .
Flunking Grades:
Research and P olicies on Retention
edited by L.A. Shepard and M.L. Smith. London:
Falmer Press.
Singer, J.D., J.S. Palfrey, J.A. Butler, and D.K. W alker. 1989. Variation in Special
Education Classification Across School Districts: How Does Where You Live
Affect W hat You Are L abeled?
merican Educational Research Journal
26:261-81.
Slavin, R.E. 1987. Ability Gro uping and Student Achievem ent in Elementary
Schools: A Best-Evidence S ynthesis. Review of Educational Research 57:293-336.
Serensen , A.B. 1970. Organizational Differentiation of Students and Educational
Opportunity.
Sociology
o f
Education
43:35576.
. 1987. The Organiza tional Differentiation of Studen ts in Schools as an
O ppo rtunity Structure. Pp. 103-29 in The
Social Organization
of Schools: New
Conceptualizations of the Learning Process edited by M.T. Hallinan. New York:
Plenum.
and M. Ha llinan. 1984. Effects of Race on Assignm ent to Ability Gro up s.
Pp. 85-103 in The
Social Context
of Instruction:
Group Organization
and Group
Processes edited by P.L. Peterson,
L.C.
W ilkinson and L.C. Ha llinan. New York:
Academic Press.
Stevenson, D., K Schiller, and B. Schneider. 1994. Sequences of Opportunities for
Learning. Sociology of Education 67:185-198.
Stroup , A.L. and L.N. Robins. 1972. Elementary School Predictors of Hig h School
Drop out A mon g Black Males.
Sociology
o f
Education
45:212-22.
Viad ero, D. 1992. Repo rt Finds Record Jum p in Special-Education Enrollm ent.
Education Week
11:19.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 123/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 124/375
Peer Social Networks and
Adolescent Career Development
Charles E. idwell
Stephen Plank
The University of Chicago
Chandra Muller
The
University
of Texas Austin
This paper explores how adolescents egocentr ic peer f r iendship net-
w ork s affect the process of career dev elo pm ent . By career de ve lop m en t, w e
m ean the processes tha t indu ce chan ge in a yo un g pe rson s be l iefs and
information about work, conceptions of self (especially as a worker) , and
expecta t ions , aspira t ions , an d decis ions abo ut edu cat ion and em plo ym ent .
Elsewhere in this volume, Gamoran, in agreement with Elder (1995) ,
arg ue s that th e life co urse can be con cep tua lize d fruitfully as a series of
transit ion s that va ry from one life to ano ther , rathe r than as a f ixed seq uen ce
of stages that is essen tially inv aria nt across lives. Lives can differ in th e
transit io ns they disp lay , in the onset and d ura tion of these transit ions , and
to som e degree in the sequences tha t they form. O R and takes the same view
of the l ife course and shows us how li t t le we know about the processes
through which l i fe course t rans i t ions and the ir var ia t ion come about—
fundam enta l ly , the contr ibut ions of hu m an agency and social ins t i tu t ions
to these processes an d th e relat ion ship s of the voli t ional and insti tutio nal in
lives bei ng l ived. M ortim er add ress es these m atters with specific reference
to occup at ional a t ta inm ent . She advo cates a socia l psychologica l app roac h
in which occupat ional h is tor ies are analyzed as outcomes of individual
effor ts to realize values and motives within a social s tructure of opportu-
ni t ies an d cons tra ints .
O ur f indings will ad dr ess cer tain of these issues. They will pro vi de
evidenc e of w ay s in which yo ung p eop le s kno w ledg e and bel iefs a bou t
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 125/375
1 8
w ork , their par t icipa tion in high school, an d their edu cation al and occ upa
tional plans vary with the form of their school-specific egocentric peer
fr iendship networks and with the beliefs and activit ies that character ize
these groups of f r iends.
With varying degrees of se l f -awareness and fores ight , young people
sha pe an d direc t the i r l ives according to w ha t they kno w and bel ieve abo ut
them selves and abou t the i r imm edia te and future prospects . W e conceive
of fr iendship n etw ork s as am on g the pr im e socia l s i tua t ions wi th in wh ich
yo un g pe op le s l ives unfold . O ur f indings will give som e indica tion of these
voli t iona l elem ents in adolescen t career de ve lop m en t. The y will also
sug ge st that the pee r netw ork s that form in the school are small syste m s of
com m unic at ion and interpersonal inf luence tha t affect y ou ng pe op le s
occupational knowledge and beliefs , their conceptions of themselves, and
their plans for education and work.
W e wil l cons ider ho w th e no rma t ive and behav iora l content a nd the form
of h igh school s tud en ts pee r ne two rks , a long wi th the r e sp on den ts ow n
centrali ty w ithin th em , affect their par t ic ipatio n in the every da y activit ies of
school and the ir longer - term kno w ledge , be liefs, and p lans abou t edu cat io n
an d jobs. W e wil l pro vid e evidence tha t these netw ork s med ia te re la tion
sh ips be tw een you ng pe r son s ca ree r deve lo pm ent and the b roa der ins t i
tutional contexts of school and, to an extent, family.
Our s tudy follows a long tradit ion of research on s ignif icant others
(friends, pare nts , teachers) as m ed iatin g actors in proce sses of ed uca tion al
an d o ccupat ion al s ta tus a t ta inme nt (e.g ., Alexan der an d C am pbe l l , 1964;
A lex an de r, Eck land , an d Griffin, 1975; H aller, 1982). This wo rk ha s h ad a
seminal inf luence on our thinking about the formation of educational and
occup at ional p lans and asp ira t ions wi thin the school context, by ope ning
u p the black bo x of the school. M ost of this w ork h as been o n the hig h school,
and i t has shown the importance of interpersonal relat ionships as l inks
bet w een wh ere s tud ents are located in a school s social an d mo ral ord er and
what they think about the i r educat ional and occupat ional prospects and
how they evaluate them.
Ho we ver , there is m ore to do. This research has not cons idered ho w
edu cat iona l or occupat iona l p lan s and aspira t ions may form in the context
of beliefs and info rma tion ab ou t w ork . In ad di tio n, it is essen tially resea rch
on dyads , in which proper t ies of ne tworks are s imple sums of dyadic
prop er t ies . For exam ple , in th is app roach, one mig ht sum the occup at ional
pla ns of a s tu de nt s s ignif icant othe rs to m ea su re interperso nal inf luence on
the s tude n t s ow n asp ir a t ions . Ho wev er , dya ds usua l ly a r e em be dd ed in
structures of social t ies , and these network s tructures can substantial ly
s t rengthen or weaken dyadic inf luence on what people th ink and do.
W he ther one thin ks of social ne tw ork s as in fact com po sed of dy ad s a nd
cliques (e.g. , Friedkin, 1993) or of structurally equivalent positions (e.g. ,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 126/375
1 9
Burt, 1992), these networks must be treated as multifunctional - social
structu res for com municating information, arrays of sentimental bond s tha t
ground influence and persuasion, and intimate arenas for self-other refer
ence that enable com parative self-evaluation. We will present findings
about the first two of these functions—how access to information affects the
range and accuracy of occupational knowledge, how exposu re to persua
sion affects values, beliefs, and participation in school, and how both
information and persuasion affect educational and occupational ambition
indirectly, via their more proximate d irect effects on school pa rticipation.
Concepts nd Propo sitions
Our conceptua l framework is based on Simmel s (1950:40,1971:41-140)
analysis of the elements of prim ary social relations, especially h is distinc
tion between their form and content. It has evolved dur ing the process of
data analys is, so that the findings that w e w ill report are no t strict tests of
a priori hypotheses.
Our argument rests on three assumptions about the consequences of
egocentric social netw orks for what people believe, know, and
do.
The first
assum ption concerns network form.
We
construe netwo rk form broad ly, to
include both properties of the network structure and properties of the
location of indiv idua ls w ithin
this
structure.
We
will
be
concerned w ith two
aspects of structure—network density and network closure. Density de
note s the prop ortion of possible ties in a netw ork that are in fact observed.
Closure refers to the rate at which a netw ork s m emb ers form ties w ith
person s outside the network. We will consider one aspect of individual
location in a netw ork. This construct is centrality—the ratio of the rate at
which a member (ego) receives interaction from others in the network
(alters) to the aggregate rate at which the alters receive interaction from ego
and all other alters.
We assume that network form can affect beliefs, knowled ge, and behav
ior through its consequences for communication and persuasion. With
respect to comm unication, the denser a network, the greater its com m uni
cative efficiency should be because as density increases, the number of
incomplete p athw ays (communication blockages) in the network decreas
es. he
mo re closed
a
network,
the
smaller the num ber of information items
that can enter
the
netwo rk because
as
closure increases,
the
num ber of entry
points from outside the network decreases. A s
a
result, the more closed the
network, the more information redundancy it should contain. Hence,
relatively few items of information are acquired by any mem ber, although
what is acquired may be learned very well. Finally, the more central a
m em ber s location in a network, the greater the num ber of information
pathways that reach this person. Therefore, the more central the member,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 127/375
11
the greater the number of items of information that this person should
receive.
W ith respect to persuasion,
we
expect the prim ary persuasive effect of an
interpersonal tie to occur as a result of the streng th of the sentimen tal b ond
in the dyad. Density
is
a reasonable measure of
the
aggregate strength of an
indiv idua l s ties to the mem bers of his or her gro up. Therefore, w e expect
the mem bers of a network to experience more persuasion in a netw ork the
greater its density. Closure may have some reinforcing effect on ne twork
persuasio n by raising the salience of the netw ork in its m em bers eyes.
Ou r second assump tion concerns netw ork content. Neither comm unica
tion nor persu asion is content-free. In the one case, information is transmit
ted and , in the other, there is an effort by one mem ber of a dyad to brin g the
other closer to some stan dard of belief or condu ct. Therefore, we assum e
that w hat a netw ork s memb ers know , believe, and d o w ill vary directly
with the aggregate level and the distribution of knowledge, belief and
beh avio r in their egocentric netw orks . About the distribution of this
content, information diversity should increase the range of any member s
kno wledge . Diversity of beliefs and behav ior should increase the norm a
tive and behavioral options open to mem ber and consequently reduc e the
likelihood of compliance with any given norm or behavioral pattern.
To some degree any such relationships will result from homophilic
selection. Nevertheless, they should also arise from com munication or
influence processes within existing network s, often
to
substantial degree.
In fact, the most interesting relationships that we will explore will be
contingent relationships between netwo rk content and network form. The
covariation of network content and individual knowledge, beliefs, or
conduct should be strengthened or weakened by variation in the p roperties
of netw ork form that we have just discussed. These contingent effects will
appear in our findings as statistical interactions of measures of network
content and form, and we expect them to be the primary ways in which
network form is associated with individual level measures of knowledge,
belief and behav ior. There may be one exception to this propos ition. If
efficient network communication and relatively open network boundaries
increase the volume and diversity of information in a network, network
density and closure may have direct effects on the accuracy and am ount of
information that members acquire.
O ur third assum ption is about the time horizons of adolescent s peer
relationships. We assume that by comparison w ith m any ad ult social
networks (e.g., colleague relationships at work, ties among neighbors or
kin), these time horizons re short. Among young people, network m ember
ship and activities tend to change frequently with age and changing
situations, so that the primary interests found in these networks are rela
tively transient (Hallinan and Tuma, 1978; Epstein, 1983). Consequently,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 128/375
I l l
FIGURE
6 1
The Career Development Process
Network
Proximal
istal
we expect the direct effects of our netw ork variables (measures of con tent
and the interactions of content and form) to be limited to behavior and
knowledge and to present-oriented beliefs. We expect these networks to
affect beliefs abo ut the longer term, such as educational and occup ational
plans,
only indirectly via their direct effects on the more proximal beliefs
and on knowledge and conduct.
These expectations are sum m arized in Figure 6.1. In this diagra m, the
relationships betwe en netwo rk form and co ntent and longer-term plan s are
me diated by the youn g p erson s cu rrent sets of beliefs, jo knowledge, and
behavio r in school. These proximal variables are themselves affected by the
interaction of netw ork form and content, with an ad ditional m ain effect of
netw ork form on job know ledge.
Figure
6 1
contains no exogenous variables. We hav e not m ade pre dic
tions concerning relationships between netw ork form or content and such
exog enous variables as the subject s own ascriptive traits, his or her socio
economic back grou nd, the size or studen t compo sition of the high school,
or the subject s h igh school curricular track, year in school, or gra de po int
average. These relationships are not our primary interest.
ata and Procedures
esearch sites and samples
Our findings come from the first year of a multi-wave study of the
adolescent a nd early a dult life course. The primary focus of this study is on
career develop m ent from the preadolescent years into the early years of full-
time labor force participation. The principal contribution of this stud y w ill
be made by analyzing developmental relationships between the subjects
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 129/375
2
evolving l ife courses and s i tuational change at the immediate level of
interpersonal social relat ions and at higher levels of social organization.
(On the dis t inc t ion between s tandard longi tudinal analys is and develop
m en tal life cou rse analysis , see Sam pso n an d La ub, 1995.) N eve rthel ess , a
pre lim ina ry exam ination of the firs t yea r s cross-sectional f inding s— here
concerning associations between l ife course s tates and interpersonal t ies—
will be instructiv e.
M ult i -wav e data are be ing gathe red in th i r teen Amer ican p ubl ic hig h
schoo ls and, for each hig h school, tw o feeder mi dd le schools . In each of
these high schools , the s tudy drew samples at random of the 1992-93 10th
an d 12th grad e s tud en t cohorts an d in each of the m idd le schoo ls of the 6th
an d 8th grade cohor ts. The sampled s tud ents are to be followed w ith an nu al
w av es of m ea su rem en t over at least f ive years ( including tho se subjects w h o
leave the original research sites).
These schools are located in twelve communities ( in one of which two
hig h schools and the cor resp ond ing m idd le schools are par t ic ipa t ing in the
s t u d y ) .
These communi t ie s va ry wide ly in economic and popula t ion
character is t ics . O ur data com e from six of the thir teen hig h schools . W e
chose s ix si tes in wh ich we cam e closest to an ade qu ate e nu m era tio n of the
resp on den ts egocen t ric ne tw orks - a t the min imu m , a m edian enu me ra
tion rate of tw en ty pe rcen t. In fact, for the six schoo ls in ou r sa m pl e, th e
m ed ia n enum era tio n rate is 0.36. The m ean is 0.35, w ith a s t an da rd
de via tion of 0.23. For the seven schools w e ha ve exclud ed, the m ed ia n
en um era tio n rate is 0.09; the me an is 0.13, w ith a s ta nd ard dev iation of 0.15.
Although we will analyze data f rom only s ix s i tes , these high schools
dif fer markedly in socioeconomic composi t ion and cur r icular emphas is .
Tw o of the schools are those from the sam e large m id w es ter n ci ty. H ere ,
Magnet High School recrui ts s tudents c i ty-wide , and, because there is a
large appl ican t po ol , i t is high ly selective academ ically. Stu de nts are cho sen
for this school on the basis of test perform anc e, teach ers rec om m en da tio ns,
an d interview s. M agnet H igh enrol ls abou t 900 s tud ents , of wh om som e 60
per cent are white and another 30 per cent Afr ican-American. The s tudent
bo dy is dive rse socioeconomically. Ab ou t 90 per cent of its s tu de nts take at
leas t one Adv anc ed P lacement college pre para tory co urse , an d the school
se nd s abo ut three-fourths of i ts gra du ate s to four-year colleges an d univ er
sities, a long w ith ano ther 20 per cent to two -year colleges. I ts four-year
drop-out rate is less than five per cent.
In contrast , Roosevelt High School serves a local at tendance area and
enrolls abou t 1400 stu de nts , of w ho m a bo ut 60 percen t are Afr ican-Am eri
can s tudents and 30 per cent white. About half of i ts s tudents come from
families belo w the Federal po ve rty l ine an d ab ou t 6 pe r cent are from single
pa rent families. H ow eve r , dro pp ing o ut of Roosevel t is com para t ive ly rare
(a four-y ear r ate of so m e ten per
cent).
Ab ou t 70 pe r cent of i ts gra du ate s go
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 130/375
3
to college, of whom about 50 per cent enroll in two-year insti tutions.
Forest View High School serves an inner-city s tudent body in one of
A m eric a s largest cit ies . Forest View h as 1,200 stu de nts , of w h o m 60 p er
cent are H ispa nic , 15 pe r cent Afr ican-Am erican, an d a no the r 15 pe r cent
no n-H ispa nic w hite . Som e 90 pe r cent of these s tu de nt s are f rom be low the
pov er ty l ine an d abou t three- four ths f rom s ingle par ent hom es . The four -
year dro p- ou t rate in this school is 40 per cent, an d of the grad ua tes abo ut
20 per cent at tend four-year colleges and another 20 per cent two-year
schools .
Middlebrook High School is located in an aff luent suburb of a prosper
ous East Coast ci ty. The high school dis tr ict is somewhat diverse occupa-
t ionally an d e thnica lly , bu t the pop ula t io n is pred om ina nt ly wh i te and is
com pose d mainly of upper - levelprofess ionals an d m ana gers . M iddleb rook
enrolls 1,700 stu de nt s . A bo ut
7
pe r cent of these s tud en ts are w hite , an d the
rem ain ing s tud en ts are racially dive rse. Very few are f rom families in
pover ty .
Th is hig h school has an established tradit io n of acade mic excellence an d
sen ds abo ut three-fourth s of i ts gra du ate s to four-year insti tutio ns. I ts four-
year dro p-o ut ra te is less than one per cent . M iddleb rook ha s s t ron g
A dv an ced P laceme nt and hon ors prog ram s in each of the pr inc ipal a r ts and
sciences areas , located w ithin a s t ruc ture of s trongly bo un de d cur r icular
tracks (college p re p , gen eral , and v ocational) . Stu de nts he re enjoy an
extens ive extra-cur r icular and a thle tic pro gra m , bu t the par t ic ip ants c om e
overwhelmingly f rom the col lege prep t rack.
Gro sse C hu te H ig h School is the central hig h school of an u p p e r
m idw est ern ind ustr ial ci ty, w ith a po pu latio n of ab ou t 200,000. This ci ty is
re la tive ly prosp ero us , but it conta ins pockets of pov er ty an d ha s an un em
plo ym ent ra te of abo ut 15 per cent. I ts po pula t ion is pre do m ina nt ly wh i te .
The city s pop ula tio n co m posit io n is m irrore d in Gro sse C hu te s 1,500
stu de nts . N inety pe r cent are w hite , an d 15 pe r cent com e from families
be low the pover ty
l ine .
T he four -year dro p-o ut ra te is a b i t und er 15 per cen t,
bu t i t i s im por ta nt to rem em ber tha t m an y of these youn g people dr op ou t
to ente r jobs or to m arry . The school m aint ains a large, specialized
vocat ional pro gr am that enrol ls about a four th of the s tud ents . A noth er
fourth a re enro lled in the college pr ep cu rr icul um , bu t in con trast to those
of some of the schools a l read y descr ibed, A dv anc ed Placem ent courses are
l imited in number and scope. Many of the vocat ional graduates cont inue
in co m m un ity college. Ab out half of the college pr ep gra du ate s at te nd four-
year sch ools , the rest tw o-y ear colleges.
River H igh School serves a m idw est ern to w n of ab ou t 15,000 inh abita nts .
This tow n prov ide s service to a sur rou nd ing farming area and is the ho m e
of a number of prosperous l ight indus tr ies . River High enrol ls approxi
m ately 1,400 stu de nts , of w ho m 90 per cent are w hite . A bo ut 15 pe r cent
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 131/375
4
come from homes below the poverty l ine. Very few are f rom single parent
families. The four-year d ro p ou t rate is f ive pe r cent This hi gh scho ol offers
a comprehens ive cur r iculum, wi th an ac t ive vocat ional component and a
sma ll college pr ep pro gra m . I ts A dv an ce d Placem ent coverage is spo tty. Of
i ts gra du ates , only about 1 pe r cent at ten d four-year co lleges, wh ile 50 per
cent enro ll in two- yea r schools (most of thes e a local co m m un ity co llege). Of
the remainder , the boys tend to go directly to work. Although some of the
gir ls go to w ork , m an y choose to m arr y an d rem ain ou t of the labor force.
Measurement
In the pre sen t s tud y, we use d three sets of data, each collected in a g ro up
ques t ion naire adm inis t ra t ion dur ing the 1992-93 school year . W e con
structed measures of the respondents ' schools and families and of their
activit ies , beliefs , and educational and occupational plans f rom responses
to i tem s in a qu estion naire that replicates portio ns of the NELS:88 surve y of
American high school s tudents ( Ingels , 1990). We buil t our measure of
occupat ional know ledge from a fifteen-item ba ttery in an oth er q ue stio nn aire
that i s devoted to knowledge and images of work. This sca le measures
accuracy of kn ow led ge a bo ut a rang e of w hite an d blue collar jobs an d job-
related issues in the A me rican occu pation al s tructu re. I t do es no t cover
w ork in the informal econom y. W e descr ibe these m easu res in grea ter de ta i l
in Appendix A.
To measure a t t r ibutes of the respondents ' egocentr ic f r iendship net
w ork s , w e drew on nom inat ion s ma de by the m em bers of our sam ples . Each
of th e resp on de nt s com plete d a form in whic h he or she lis ted the n am es of
u p to fourteen peo ple in resp on se to the que stion, W ho are the f r iends y ou
usual ly ha ng arou nd w ith? In add i t ion to l is t ing the na m es , the respon
dent gave each nominee 's gender and indica ted whether the f r iend was a
neighbor , classmate, co-member of an activity l ike a band or club, or
something else (which was specif ied).
1
The samp les we re stratif ied by ge nd er an d ethnicity and , in som e of ou r
schools , by level of academic performance (at the minimum honors vs .
others ) . Th us , w e hav e data tha t can be used to form chooser -chosen
matr ices for network analysis , accompanied by other data that can be used
to der iv e m easu res of ne tw ork comp osi t ion.
We cons tructed s imple measures of ne twork s ize and s t ruc ture . Our
m eas ure of ne tw ork s ize is the nu m be r of a l ters no min ated b y the respo n
de nt f rom his or he r school an d gra de . In l ieu of a s ta nd ard m ea su re of
ne tw ork den sity, w e will use a m easu re that w e call cohesion . In ou r usa ge,
cohes ion is the propor t ion of a l te rs in an egocentr ic ne twork who nomi
nated at least one other al ter in this network as a member of their own
egocentr ic f r iendship networks.
W e decid ed to use th is cohes ion me asu re ra ther than a s tan da rd dens i ty
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 132/375
115
m easure because we discovered that the distribution of alter-alter n om ina
tions was severely skewed. Relatively few alters in any of these netw orks
nom inated m ore than one or two other
alters,
so that the ratio of possible to
observed one-way within-network ties is badly attenuated. Ou r cohesion
m easure and this ratio, which is the standard density measure, are reason
ably strongly correlated (r = .63).
Our measure of closure is the ratio of within-network ties originating
from the ne twork s members to the total num ber of ties originated by these
resp ond en ts. Finally, our measu re of centrality is som ewhat different from
the most com monly used measures of centrality. We have defined a
resp ondent s (ego s) centrality as the quotient of two
ratios.
The ratio in the
num erator is ego s observed num ber of nominations received from the
alters in the egocentric network d ivided by ego s possible received nom ina
tions. Possible nom inators are limited to persons who were nom inated by
ego and w ho , in turn , responded to our questionnaire. The ratio in the
denom inator is the average of the comparable
ratios,
calculated for each of
ego s nom inated friends (alters). Possible nom inations for each of the alters
are again limited to those from people in ego s network who respon ded to
our questionnaire.
Because our network measures are based on friendship ties to other
students within the responden t s school grade, we will lose some informa
tion about peer relations outside
the
school and grade as
a
context for career
development. However, we assume that the school grade for most high
school students encompasses the greater number of friendship choices.
Findings
Although we have made no predictions about the action of exogenous
variables, several variables that were indicators either of the respondents
ascriptive traits or social origins or of their curren t school settings proved
to be substantially correlated w ith endogeno us and depende nt variables in
our mode ls. Sets of these variables we re treated as exogenous in our
m ode ls. Therefore, each of the series of mode ls to be described be low
consists of four blocks of variab les. (See Append ix A for detailed informa
tion about these variables.) The first three blocks include the exogenous
variables. Block 1 contains the measu res of ego s gender and level of
parental education. In one series of models, this block also includes mea
sures of parenta l occupation and ego s own educational expec tations. Block
is a series of dum m y variables for ego s school site, serving as a proxy for
such school attributes as the composition of the student body and the degree
and kind of curricular specialization. Block 3 locates ego in this school,
either
in
relation
to
the distribution of student attainment—the respo ndent s
grade point average (GPA)—or in relation to the formal structure of the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 133/375
216
school—year in school and self-reported high school track placement.
These blocks of exogenous variables are fol lowed by measures of the
conten t of ego s ne tw ork
belief,
in format ion , or behavior) , measures of
ne tw or k form, and term s for the interact ion of conten t an d form. W e wil l
repo r t the Ord inary Leas t Squares (OLS) eva lua t ion of these m od els .
A lth ou gh ou r data are cross-sectional , w e ha ve som e basis for inferences
about the re la t ive importance of se lec t ion versus communica t ion and
persu as ion in the re la t ionships we wil l observe be tween ne tw ork form and
conten t and ego s m easu red be l ie fs, behav ior , and p lans . W hen se lec t ion
effects are s trong, there should be correspondingly s trong correlat ions
b e t w e e n e g o s
belief,
behav ior , or p lan an d the corre spon ding m ean for the
a l te rs in the ne tw ork. In th is case, exp and ing a m od el to inc lud e m easu res
of netw or k s tructu re, like cohesion or central i ty, and term s for the interac
t ion of these variables with network content , should add l i t t le to the
ex pla ine d varia nce . At the sam e tim e, the size an d statistical significance of
the parameter es t imates for the f irs t and third blocks of exogenous vari
ables ,
that is , the m eas ure s of ego s ow n trai ts , sho uld be redu ced . To the
exte nt tha t these tw o criteria are satisfied, w e will infer that selection effects
are s tronge r than those of com m unica t ion or persuas ion . To the extent tha t
adding te rms for ne twork s t ruc ture and s t ruc ture-content in te rac t ions
increases a mod el s predic t ive pow er an d to the extent tha t p ara m eter
est imates for the f irs t and third blocks of exogenous variables are not
reduced in s ize and s ignificance, we wil l infer that communication or
influence effects were relatively strong.
articipation in school
Tables
6.1
an d 6 .2 presen t the eva lua t ion of m od els tha t pre dic t beh avio r
in school. Table 6.1 rep orts the f indings for the nu m be r of ex tra-curr icular
activi t ies enga ged in (other than spo rts) , w hile Table 6.2 con tains the results
for school trouble (our measure of the severi ty of the respondents
self-
reported infractions of school rules).
2
Th ese tables pre sen t a sha rp contrast , acco rding to the foregoing reaso n
ing . Extra-curr icular par t ic ipa t ion appe ars to hav e been subs tant ia l ly
inf luenced wi th in ego s ne tw ork. Ho we ver , a l thou gh we f ind a s igni ficant
assoc ia tion be tw een eg o s school t rouble score and the a l te rs m ean score ,
this relat io nsh ip ap pe ars to ha ve arisen chiefly from the selective form ation
of peer ties.
In Table
6.1,
M odel I rep orts the effect pa ram ete rs for a m od el in w hich
the predictors are our three blocks of exogenous variables . The results are
straigh t-forw ard. Fem ales are m ore l ikely tha n m ales to be invo lved in the
non-spor ts ext ra -curr iculum, as a re respondents wi th be t te r educa ted par
ents . By com pariso n wi th Gran d Chu t te resp on den ts ( the re ference ca tego
ry),
resp on den ts f rom M agnet H igh School show an unus ual ly h igh leve l of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 134/375
7
TABLE 6.1 Models of Ego's Non-Sport Extracurricular Participation
Intercept
Female
Parent's E ducation
Forest View
River
Middlebrook
Roosevelt
Magnet
GPA
Senior
College Prep Track
Voc/Tech Track
Mean of Frien ds' # of
Non-Sports
Cohesion
Cohesion*Mean of Friends'
# of N on-Sports
Centrality
Centrality*Mean of Friends'
# of Non-Sports
Adjusted R
2
I
-0.51
0.34***
0.10***
-0.11
-0.46*
-0.35*
0.27
0.59***
0.41***
0.43***
0.47***
0.42
0 234
II
-0.59*
0.34**
0.09**
-0.06
-0.39*
-0.20
0.24
0.45*
0.37***
0.34**
0.41**
0.43
0.18***
0 249
III
-0.28
0.37***
0.09**
-0.04
-0.33
-0.14
0.28
0.40*
0.36***
0.36**
0.37**
0.46*
-0.05
-1.36**
0.94***
0 265
IV
-0.45
0.35**
0.09**
-0.05
-0.35
-0.18
0.26
0.41*
0.37***
0.33**
0.40**
0.42
0.08
-0.18
0.11*
0.251
n = 65 *p<.0 5 **p<.01 ***p<.001
a
See Appendix A for descriptions of dependent and independent variables.
extra-curricular participa tion, while those from River High, the small tow n
school with
a
restricted extra-curriculum, an unusua lly low level. It
is
more
surprising to find that responde nts from Middlebrook H igh also are com
parative ly less active participan ts in the extra-curriculum. How ever, o ther
of our d ata show that the Middlebrook comm unity w as distinctive in the
range of non-school activities available to high school youth, and many
M iddlebrook students take part in these program s. Finally, getting good
grades, being in the college prep rather than the general track, and being a
senior rather than a sophom ore were each positively associated w ith extra
curricular participation.
In the remaining three models, we introduce measures of the egocentric
netw orks. Ad ding these variables adds appreciably to the explained
variance, while the effect parameters for the first and third blocks of
exogenous variables are essentially unaffected. It is interesting to observe
that the significant associations between school site and extracurricular
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 135/375
8
TABLE 6.2 Models of Ego s School Trouble
Intercept
Female
Par en t s Educat ion
Forest View
River
Midd leb rook
Roosevel t
Magne t
G P A
Senior
College Prep Track
Voc /Tech Track
Mean of Friends School
Trouble
Cohes ion
Cohesion*Mean of Fr iends
School Trouble
Central i ty
Cent ra l ity*Mean of Fr iend s
School Trouble
Adjusted R
2
I
7.87***
-0.38
-0.07
0.32
-1.88***
-0.09
-0.61
-1.46***
-1.09***
0.37
-0.20
0.82*
0.202
II
5.87***
-0.27
-0.08
0.11
-1.37***
0.03
-0.49
-1.03**
-0.88***
0.30
-0.10
0.73
0.32***
0.253
III
6.26***
-0.24
-0.08
0.08
-1.28***
0.07
-0.56
-1.00**
-0.85***
0.27
-0.09
0.69
0.26***
-1.68
0.19
0.256
IV
6.00***
-0.22
-0.07
0.11
-1.32***
0.04
-0.51
-0.98**
-0.87***
0.30
-0.09
0.68
0.31***
-0.18
0.01
0.252
n = 710 * p < .0 5 **p<.01 ***p<.001
a
See Append ix A for descriptions of dependent and independ ent variables.
par t ic ipa t ion are to a m od es t degree med ia ted b y proper t ies of the respo n
d e n t s
egocentr ic ne tworks .
In M od el II , w e f ind a substantial association bet w ee n th e me an level of
a l t e r s an d ego s par t ic ipa t ion. W hen we int roduce the me asures of ne tw ork
form— cohesion in M ode l III and ego s centrali ty in M od el IV—w e f ind
signif icant interactions of network form and content.
3
In eac h case, the
association betw ee n the aggreg ate activity level of the al ters in the n etw or k
and ego s act iv i ty becom es s t ronge r the m ore cohes ive the net w ork an d the
m ore central eg o s location in i t.
4
The param eter es t imates in Table
6.2
for the pre dic tion of schoo l trou ble ,
sug ge st a stro ng er influence of selection effects t ha n tho se of the prio r table.
In M od el I , the results again are sensible. The m ore tro ub le-pro ne resp on
de nts are least likely to be found in M agn et or River Hi gh an d, within their
schoo ls , m os t likely to be in the vocational track. They ten d to hav e lo w
grades .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 136/375
219
When the network measures enter the models (Models II-IV), the ex
plained variance increases, but the only significant network correlate of
ego s school trouble is the level of his or her frien ds school troub le. The
significant coefficients in the third block of exogenous variables, for GPA
and vocational track placement, now weaken and, in the latter case, lose
statistical significance. Ne ither of the mo deled s tructura l variables has a
significant main effect, and neither interaction term (content by form) is
significant. If ou r reasoning is correct, we can infer that involvem ent in
school troub le to a substantial degree preceded friendship formation.
fo knowledge
Table 6.3 presents the parameter estimates for five models that predict
the resp on den ts scores on our job knowledge scale. Job know ledge should
be especially sensitive to exposure to information about jobs. For this
reason, we constructed the mo dels to represent the resp ond ent s involve
ment in four information environ ments: the family, the school, locations
within the formal structure of the school, and the egocentric peer network .
Model contains the blocks of exogenous variables. This m odel has been
exp and ed to include a series of dum m y variables that classify the occupa
tion of whichever of the respo nd en t s p aren ts is the fam ily s principal
bread w inner. These categories group the occupations according to sub
stance as well as standing—more complex and less complex professional
jobs (Professional II vs. Professional I), managerial, lower white collar,
laborer, and other blue collar. In this mode l, lower wh ite collar is the
omitted
category. We
assu m e that these job categories are reasonably valid
indicato rs of the focus and ran ge of job information available in the family.
Given the con tent of the job knowledge scale, one m ight expect the children
from families of m idd ling occupational stand ing
to score
unusually well, on
the assum ption that the content of the parental occupation was reflected in
the content of job information curren t in the family. Ho we ver, scores on the
job know ledge scale depen d substantially on a range of know ledge, and a
family s rang e of information might be expected to increase in ro ugh
approximation to increases in occupational standing, as a function of
differences in families human capital resources.
Model introduces the network m ean on the depe nde nt variable. Model
III exten ds the representation of netw ork con tent to include the responden t s
report of the principal c urrent interests and activities of the alters in his or
her network—the academic and social factor scale scores—to determine
whether the chief activities and interests of this group of friends would
affect the availab ility of job information in the netwo rk and the respondent s
sensitivity to this information. Models IV and V introduce , respectively,
netw ork cohesion and ego s centrality, each accompanied by the co rrespond
ing interaction term.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 137/375
12
TABLE
6.3
Intercept
Female
Paren t ' s Educat ion
Ego 's Educat ional
Expectat ions
Mode l s
of
Ego's
I
-0.58
0.09
0.05
0.06
II
-0.58
0.09
0.05
0.06
fob Knowledge
III
-0.80
0.10
0.05
0.08*
Score
IV
-0.60
0.11
0.05
0.09*
V
-0.53
0.10
0.06
0.08*
Professional I
Professional
II
M a n a g e r / A d m i n i s t r a t o r
Laborer
Other Blue Collar
Forest View
River
Midd leb rook
Roosevel t
Magne t
-0.13
0.28
0.17
0.10
0.25
-0.40
0.18
0.20
0.13
0.65**
-0.13
0.27
0.17
0.10
0.25
-0.37
0.18
0.19
0.12
0.62**
-0.15
0.22
0.19
0.07
0.23
-0.28
0.23
0.17
0.20
0.65**
-0.13
0.21
0.18
0.08
0.21
-0.33
0.27
0.13
0.20
0.57*
-0.16
0.21
0.15
0.04
0.22
-0.33
0.29
0.16
0.20
0.55*
GPA
Senior
College Prep Track
Voc/Tech Track
M e a n
of
F r i ends '
Job Knowledge
Academic Fac tor
Social Factor
Cohes ion
Cohesion*Mean
of
Friend s '
Job Know ledge
Cent ra l i ty
Cent ra l i ty*Mean
of
Friends
Job Knowledge
Adjusted
R
2
0.17*
0.36**
0.43**
0.00
'
0.127
0.17*
0.35*
0.42**
0.00
0.03
0.125
0.19*
0.36**
0.45**
0.07
0.03
-0.24***
-0.01
0.141
0.19*
0.34*
0.45**
0.06
-0.15
-0.23
0.01
-1.03
0.77**
0.150
0.18*
0.37**
0.43**
-0.01
-0.18
-0.23***
0.01
-0.24*
0.22**
0.155
n
=
552
p < .05
'
p<.01 p<.001
See Appendix A for descriptions of dependent and independent variables.
To the extent that we have been able to measure the information environ-
ment of the respondents' families, the parameter estimates for Model I
suggest that these families were not effective information sources for these
young people. Neither the type of parental job nor the level of parental
education is a significant predictor. Nonetheless, we had expected parental
jobs to affect both the focus and range of job information in the family.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 138/375
121
M oreov er , one wo uld expect the var ie ty of job informat ion , an d pe rha ps the
frequency of family discu ssion ab ou t such m atte rs , to be a pos i t ive co rrelate
of parenta l educa t ion .
There
s
evid enc e th at the school i tself is a m or e effective e nv iro nm en t for
in fo rming o ur re spo nde n t s abou t jobs . Respo nden t s f rom M agne t H igh
scored excep tionally w ell on the job kno w led ge scale . In contrast to each of
the o ther schools , inc luding the curr icular ly demanding Middlebrook
H igh, Ma gne t Hig h enrol ls a s tud ent b od y of h ighly d iv erse soc ial o r ig ins
w i th ou t be ing in tens ive ly t racked. Th us , in M agne t High informat ion abo ut
a d ivers i ty of jobs and work worlds could be communica ted f rom one
s tud en t to ano the r w i thou t encoun te r ing the s t ruc tu ra l b lockages induced
by a t racked curr ic ulum . Tracking, by contras t, un do ub ted ly imp oses
subs tant ia l l imi ts on s tudents abi l i ty to in te rac t wi th o thers whose back
grounds and experiences differ from their own.
Within the school , s tudents who are h igh performers score be t te r than
others , as do seniors compared wi th sophomores ( the la t te r f inding no
doubt reflect ing maturat ional and experiential changes in the l i fe course).
Of gre ater interest , pe rh ap s, w e see that the sam e track bo un da rie s that l imit
and segregate s tudent in te rac t ion , thereby evident ly def in ing d is t inc t ive
job information environments . They do so irrespective of the school in
which they are located. The college prep track, compared with ei ther the
gen eral or the vocation al or technical t rack seem s to confer an inform ation
adv an tage . Th i s adv an tag e p resum ably has s eve ra l sources—the range a nd
explici tness of information treated in courses , in face-to-face interact ion
with teachers , counselors , and fel low students , and in the context of those
extra-curricular act ivi t ies , l ike debate or publicat ions, that draw their
part ic ipants chiefly from the college prep track.
In Mo del II , w e intr od uc e the aver age level of the ne tw ork al ter s job
knowledge scores and in Model III the two factor scores in which the
resp on den t descr ibes the degree to wh ich h is or her ne tw ork of f r iends i s
orie nted in interests and act ivit ies to thin gs academ ic or m or e heav ily in the
rea lm of ado lesce nt sociabil i ty. M od el III pre sen ts a pe rple xin g finding, to
which w e wi l l turn in a m om ent . No te f irst tha t s imply ha ving f r iends w ho
are well inform ed ab ou t w or k do es not in i tself seem to confer an informa
t ion advantage on the respondents (Model I I ) . However , the parameter
est im ates for the conten t-form interact ion ter m s in M ode ls IV an d V are
significant. W he n these interac tion s are plo tted (see Fig ure s 6.2 an d 6.3), the
plots show tha t w he n ne tw ork coh es ion is low or w he n ego s loca tion in the
netw ork is re la tive ly per iphera l , the re la t ionship be tw een the ne two rk job
score m ea n an d ego s me an is essentia lly zero . H ow ever , the m ore cohes ive
the ne tw ork or the m or e centra l ego s locat ion , the m ore s t rongly pos i t ive
th is re la t ionship becomes .
These f indings are consis tent with our notion of information environ-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 139/375
122
FIGURE 6.2 Mean of
Alters
Job Know ledge Scores,
Ego s
Job
K nowledge Score,
Cohesion
2.0 T
1 1.0
0.96
3
1.90
1.26
o
— O- Cohesion = .1
— •-Cohesion = .6
0.0
Mean of Alters Job Know ledge Scores
FIGURE 6.3 Mean of
A lters
Job Know ledge Scores,
Ego s Job Know ledge Score, Centra lity
2.0
a
§
1
1.46
1.0
1.10
o
tic
—D—
Centrality
.5
- • - Centra lity = 2
0.0
Mean of Alters Job Knowledg e Scores
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 140/375
3
m ents. They suggest that, along with the school and the track within the
school, the egocentric netw ork comprised
a
series of paths for co mm unicat
ing information abou t the work w orld. In this interpretation, the more
cohesive the netwo rk, the more com plete the set of path s and the fewer the
com munication b arrie rs. Similarly, the more central ego s location, the
greater his or her relative expo sure to incoming
messages.
In short, both of
the interaction terms suggest conditions affecting the efficiency of the
egocentric network as an information system.
The greater efficiency of the more cohesive networks may also have
arisen from the greater salience or persuasiveness of information coming
from others w hom ego trusted (as a function of positive sentim ent), while
our finding concern ing centrality m ay indicate effects not only of comm u
nication, but also of status (e.g., the most informed in a group of the
informed thereby gaining a central position). Our data do not allow us to
disentangle the specific mechanisms involved in these associations, but
communication almost certainly would prove
to
be prominent among them.
In any event, selection mechanisms to not appear to be the primary
source of the association of netw ork prop erties and
job
know ledge scores.
In the m ode ls of Table 6.3, the block of netw ork variables ma kes its ow n
contribution to predictive power, while adding m easures of netw ork struc
ture to M odels IV and does not alter the significant p aram eter estimates
in the third block of exogenous variables.
It is difficult to und ersta nd the negative sign of the coefficients for Ego s
Academic Factor Score found consistently in Models III through V. The
nega tive coefficient does not app ear to be artifactual. If the more academ i
cally oriented n etw orks are those in which the greater am oun t and diversity
of information circulates, an intuitively reasonable expectation, then this
coefficient should b e positive. Note that once this factor score enters the
mode l, the coefficient for the respo nde nt s own educa tional expectations,
which is positive, becomes statistically significant, suggesting that friends
academic orientations depressed the job know ledge scores of even the more
academically oriented respon dents.
This finding might be interpreted as indicating a lack of information
diversity in the more academically oriented networks, as a result of some
distinctive narrow ing o r focusing of interests or experiences. How ever, a
fairly wide range of interest and experience indicators available in ou r da ta
did not show m arked differences between the more and the less academ i
cally oriented ne twork s. If any thing, the members of the more academ i
cally-oriented gro up s of friends w ere somewh at more likely than others to
participa te across a spec trum of information-giving activities. Perhap s
subtle interpersonal com parison processes were at work that we could n ot
discern. In any event, this finding requires more investigation and , for the
moment, remains a puzzle.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 141/375
4
ducational and occupational plans
Table 6.4 reports our f indings for the predict ion of the respondents
edu cat ion a l p la ns . We obta ined som ew hat s imi lar f indings for the predic
t ion of expe cted occ upatio nal a t ta in m ent, looking ahe ad to age 30, bu t th e
pre dict iv e po w er of these m od els is very small . T herefore, we wil l l imit ou r
discussion to the predict ion of educational plans .
W e in t roduce the m ean of the a l te rs educa t ion a l p lans in M odel
II .
Then ,
fo l lowing the conceptua l f ramewo rk tha t w as sum m arize d in Figu re 6 .1, w e
in t rod uce three mea sures of the resp on den t s proxim al in-school beha v
ior—the number of extracurricular act ivi t ies (other than athlet ics) , the
h ou rs spen t in these act ivi t ies , an d the school trou ble score. Ex tra-curricula r
h ou rs an d school troub le pr ov ed to be s ignificant pre dicto rs an d first
app ea r in M ode l
III.
In M ode l IV, w e add the resp on den t s depic t ion of the
acad em ic and social orientat ions of the fr iendship gr ou p and in the rem ain
ing mod els the measu res of ne tw ork cohes ion, ego s cent ra l i ty , and the
accompanying content - form in terac t ion te rms .
For the most part , these f indings replicate those of earl ier educational
a t ta inm ent research , sho w ing the d ist inc tive impo rtance of soc ia l or ig ins ,
school at ta in m ent, t rack place m ent, and the beliefs, act ivi ties , an d asp ira
t ions of significant others a m on g a s tu de nt s peer s for the s tu de nt s ow n
edu cat ion a l p lans . H ow eve r , w i th respec t to these assoc iat ions be tw een
eg o s plans a nd n etw ork con tent , the configu rat ion of coefficients and
predict ive power in this series of models suggests that s trong select ion
effects were at work.
For us , the mo re interest ing findings h av e to do w ith the app ar en t role of
extra- curricu lar part ic ip ation as a m ed iat in g variable l inking the con tent of
these s tu de nt s f r iendship ne tw ork s to the i r educa t iona l expec ta t ions . W e
fou nd , in a series of m ode ls not rep orted her e, that the nu m be r of ho ur s that
ou r resp on den ts spent in the non -spor ts ext racurr iculum w as s ignif icant ly
pred ic ted by m ean a l te r s hours of such par t ic ipa t ion , wi th a pa t te rn in the
es t ima ted m ode ls sugges t ing tha t m ore than se lec tion ha d been a t w or k to
pr od uc e th is assoc ia t ion . W e un der too k a pa th ana lys is pred ic t ing eg o s
educat iona l expec ta t ions , in which , among the endogenous var iables , the
acad em ic an d social factor scores an d al ters ho ur s are prio r to eg o s ho ur s .
This analysis show s that abo ut half of the total effect of a l ters ho u rs on eg o s
ed uc atio na l plan s is indirect , via ego s extra-curricular ho urs . M ost of th e
total associat ion of the two factor scores with the dependent variable is
direct . Th ese results pr ov ide a tentat ive indicat io n that certain n etw or k
effects on the longer-term, more dis tal outlooks of our respondents arose
from m or e pro xim al pa ttern s of act ivity (and pe rh ap s associated beliefs a nd
information) induced within the context of peer fr iendship.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 142/375
125
TABLE 6.4 M ode l of Ego s
I
Intercept 4.90***
Female 0.21
Paren t s Educa t ion
0.11**
II
3.98***
0.21
0.11*
Education
III
4.40***
0.22
0.09*
al Expectations*
IV
4.59***
0.16
0.09*
V
4.58***
0.15
0.09*
VI
3.91***
0.10
0.07
Forest View
River
Midd leb rook
Roosevel t
Magne t
G P A
Senior
College Prep Track
Voc/Tech Track
-0.14
-0.22
0.52*
-0.21
-0.04
0.35***
0.28
0.91***
-0.29
-0.07
-0.17
0.41
-0.18
-0.10
0.32***
0.25
0.85***
-0.22
-0.11
-0.19
0.46
-0.10
-0.11
0.22*
0.29
0.75***
-0.17
-0.19
-0.25
0.36
-0.20
-0.17
0.20*
0.19
0.73***
-0.16
-0.19
-0.26
0.36
-0.19
-0.17
0.20
0.20
0.73***
-0.16
-0.15
-0.36
0.40
-0.17
-0.22
0.20
0.20
0.73***
-0.07
M ean of Fr iends Educat ional
Expectat ions
Ego s School Tr oub le
Ego s Ho urs of
Ext racurr icu lars
Academic Fac tor
Social Factor
Cohesion
Cohesion*Mean of Fr iends
Expectat ions
Central i ty
Cent ra l i ty*M ean of Fr iend s
Expectat ions
Adjusted R
2
0.201
0.15*
0 209
0.11
-0.05*
0.16**
0 228
0.11
-0.02
0.14**
0.26**
-0.18*
0 247
0.11
-0.02
0.13*
0.26**
-0.18*
0.07
0.01
0 244
0.18*
-0.02
0.13*
0.25**
-0.19*
0.80
-0.08
0 257
n = 5 0 3 *p<.05 **p<.01 ***p<.001
a
See Append ix A for descriptions of depend ent and ind ependent variables.
onclusion
For the most part, our findings are consistent with the conceptual
argument with which we opened this paper and with the life course
perspective that characterizes many of the other pape rs in this volum e. So
far as proximal, in-school behavior is concerned, we have observed a
systematic interplay of form and content in the egocentric friendship
networks of our respondents that strongly suggests the persuasive influ
ence of cohesive social ties on conduct, perhaps in conjunction with the
social controls of norm -setting and self-other com parison.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 143/375
126
In our findings on job knowledge an d its correlates, we have found no
less suggestive evidence of the importance of the netw ork as a communica
tion system and of netwo rk form as a set of attributes affecting the syste m s
efficiency. Striking here is the way in which this informal system app ears
to work side-by-side with the communication pathways provided by the
more formal structure s of school curricula and by local district an d school
policies governing tracking and the composition of the student body.
Ou r findings are tentative and our m easurem ent, especially of the family
context, is incomplete. However, there are suggestions in these data of a
possible reduction in the importance of the family as a source of occupa
tional information for high school youth, giving the school, in both its
formal and informal social organization, a more central place. One m us t
bear in mind the modest predictive power of this series of models, which
probably indicates unmeasured within-school communication and the
action of information sources external to our conceptual scheme. Neverth e
less, if ou r findings are borne out, they will depict the high school as an array
of formal and informal social structures that make it an im portant locus for
interventions designed
to
increase the occupational inform ation effectively
available to young people.
The friendship netwo rk as a place where information is com mu nicated
and proximal behavior and, presumably, beliefs are influenced, and how
these netw orks stand in relation to other clusters of social ties and the m ore
macroscop ic levels of social organization—formal organizations and insti
tutions—are m ajor ideas underlying our conceptual argum ent a nd major
stra nd s in our findings. Especially in our greater ability
to
predict proximal
events than more distal outcomes, we can see how involvement in netw orks
of ties to peers (and by implication to adults), rathe r than dete rm ining life
course events or outcom es, prov ides repertories of information, beh aviors ,
and beliefs that the young person can use as he or she makes decisions about
next steps in life. o doub t the richness and the content of these reperto ries
is in good m easure a m atter of institutionally ord ered oppo rtunities and
constraints—a major point of juncture between the institutional and voli
tional in the life course.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 144/375
7
Notes
The research repo rted in this chapter was suppo rted by
a
gran t from the Alfred
P. Sloan Foundation. The authors alone and not the Foundation are responsible for
the contents herein. We are grateful for the research assistance of Daniel M cFarland,
Rebecca Sandefur, a nd Jennifer Schmidt. We are no less grateful for the critical
comments of members of the Duke University Conference on Institution and
Careers, the Sociology of Education Brown Bag Seminar at the Ogbum-Stouffer
Center, University of Chicago and colleagues in the Department of Sociology,
University of Massachusetts, Amherst.
1.
In addition, the respondent reported the frequency of interaction with each
nom inee, which of them was a best friend, and which of them was a source or
object of advice. These data are not used in the present analysis.
2.
We examined five other proximal outcomes. These were absenteeism from
school, num ber of athletic extra-curricular activities, hours spent in extra-curricular
activities (athletics and o thers separately), ho urs of unassigned read ing, and hours
spent doing homew ork. We observed some effects of netw ork form and conten t, bu t
the findings are less clear-cut than those presented in Tables and 2.
3. We do not include cohesion and centrality in the same m odel because they are
highly correlated (r = .46).
4. A lthough we hypothesized some effect of network closure and also thoug ht it
important to include network size in our models, neither variable had significant
effects when these models were estimated. In addition, for some of our models
(including the series for job know ledge shown in Table 3), we suspected that there
might be a three-way interaction between cohesion, closure, and alters' mean job
know ledge. How ever, the parameter estimates for this three-way interaction were
not significant. In the light of the foregoing, we have removed network size and
closure from each of the m odels that w e discuss.
eferences
Alexander, C. No rman and Ernest Q. Cam pbell.
1964.
Peer Influences on Adoles
cent Educational Aspirations and Attainments.
American Sociological Revkw
29:568-575.
Alexander, Karl N., Bruce Eckland, and
LarryJ.
Griffin. 1975. The Wisconsin Model
of Educational Achievement. American journal of Sociology 81:324-342.
Burt, Ronald. 1992.
Structural Holes.
Cam bridge, Mass.: Harvard University P ress.
Elder, Glen H., Jr. 1995. Hum an Lives in Chang ing Societies: Life Course and
Developm ental Insights. In
Developmental Science: Multi Disciplinary Perspec-
tives edited by R. B. Cairns, G. H. Elder, Jr. and E. J. Costello. New York:
Cam bridge U niversity Press.
Epstein, Joyce L.
1983.
Selection of Friends in Differently Organ ized Schools and
Classrooms. Pp . 73-92 in Friends in
School
Patterns o f Selection and Influence in
Secondary
Schools edited by Joyce L. Epstein and Nancy Karweit. New York:
Academic Press.
Friedkin, Noah
E.
1993. A Formal Theory of Social Power.
American Sociological
Review
58:861-872.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 145/375
128
Haller, Archibald 0 .19 82 . Reflections on the Social Psychology of Status Attain
ment. Pp . 3-28 in
Social Structure and
Behavior edited by Robert M. Hauser,
David Mechanic, Archibald O. H aller and Taissa S. H auser. New
York:
Academic
Press.
Hallinan, Maureen T. and Nancy B. Tuma. 1978. Classroom Effects on Changes in
Children's Friendships.
Sociology o f Education
51:270-282.
Ingels, Steven J. Samir Y. A braham , Rosemary Karr, Bruce D. Spencer and M artin
D. Frankel. 1990.
NELS:88 Base Year Student Component Data File User s
Manual.
Washington, D. C : National Center for Educational Statistics.
Sampson, Robert J. and John H. Laub. 1994. A Life Course Theory of Cu mulative
Disad vantage and the Stability of Delinquency. In
Developmental Theories
of
Crime and
Delinquency:
Advances in Criminological
Theory
Vol.
6
edited by Terence
P.
Thornberry. New Brunswick, N.
J.:
Transaction Publishers.
Simmel, Georg.
1950. The Sociology ofGeorg
Simmel. Edited and translated by Kurt H.
Wolff. New York: The Free Press.
1971.
On Individuality and Social Form s. Edited by Donald N. Levine. Chicago:
University of Chicago Press.
Appendix A
escriptions an d istributions of Variables
I Descriptions of Variables
Dependent Variables
Non-Sport Extracurricular
School Trouble
Job Knowledge Score
Ranges from 0 to 9, based on Participation
participation in each of nine categories of school-
based , non-athletic activities.
Ranges from 0 to 24, based on degree of trouble
experienced in each of six categories of school-
related disciplinary problems.
Rasch score constructed from twelve true-false items
about a wid e range of jobs and work-related terms.
The range of items includes topics especially rel
evant to blue-collar and craft jobs (e.g., An app ren
tice is a new worker wh o is assigned to learn a trade
from a more skillful worker ), as well as topics
relevant to expert professions (e.g., Most lawyers
spend their wo rking days in court room s ). The
informal econom y is not represen ted in the scale.
Educa tional Expectations From question, As things stand now, how far in
school do you think you will get? Response cat
egories are: l=les s than h.s. graduation; 2=h.s.
graduation only; 3=less than 2 years of vocational,
trade, or business school; 4=2 years or more of
vocational, trade, or business school; 5=less than 2
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 146/375
229
ndependent Variables
Femaie
Paren t ' s Educat ion
Professional I
Professional II
M a n a g e r / A d m i n i s t r a t o r
Laborer
Other Blue Collar
Forest View
River
Midd leb rook
Roosevel t
Magne t
GPA
Senior
College Prep Track
Voc/Tech Track
M ean of Fr iend s '
Number o f Non-Spor t s
Mean of Fr iends '
School Trouble
years of college; 6=2 or m ore y ear s of college; 7=B. A.;
8=M.A., or equivalent ; 9=Ph.D., M.D., or equiva
lent.
D um m y var iab le wi th m ale as the exc luded refer
ence category.
Edu cat ion level of m ore highly edu cate d p are nt (or
the one p resen t , if only one is pres ent) .
Resp onse categories are l= di d n ot f inish
h.s. ;
2= h.s.
graduat ion or GED; 3=vocat ional school , junior
college, or other 2-year school; 4=some college;
5=B.A.; 6=M.A., or equivalent ; 7=Ph.D., M.D., or
equivalent .
Series of f ive dummy variables represent ing the
occupat ion of the family 's principal earner
(father, if prese nt an d emplo yed ; othe rw ise,
m othe r). Professional I inclu des jobs suc h as
accountant and registered nurse. Professional II
includes jobs such as physic ian and lawy er. M an
ager /Adminis t ra tor inc ludes jobs such as sa les
ma nager and res tauran t m anager . L aborer inc ludes
jobs such as construct ion w ork er and farm laborer.
Other Blue Collar includes jobs such as mechanic
and beaut ician. The exclu ded reference category ,
Lower White Collar, includes jobs such as bank
teller and dental technician.
Series of f ive dummy variables represent ing five
of the six schools, w ith G rosse Ch ute as the
excluded reference category.
Ran ges from 0.5 to 4.0, wi th 0 .5 represen t ing grad es
most ly be low D and 4 .0 represen t ing mo st ly
A's .
Du m m y va r iab l e wi th soph om ore a s t he exc luded
reference category.
Pa i r o f dummy var iab les wi th
general t rack as the excluded reference category.
See descript ion of dependent variable, above.
Mean based on f r iends for whom non-miss ing re
sponses w ere ava i lab le .
See descript ion of dependent variable, above.
Mean based on f r iends for whom non-miss ing re
sponses w ere ava i lab le .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 147/375
130
Mean of Friends' Job
Know ledge Score
Mean of Friend s'
Educational Expectations
Academic Factor
Social Factor
Cohesion
Centrality
See description of dependent variable, above.
Mean based on friends for whom non-m issing re
sponses were available.
See description of dep ende nt v ariable, above.
Mean based on friends for whom non-m issing re
sponses were available.
First factor (eigenvalue=3.83) deriv ed from a factor
analysis of eleven items following the question,
Among the friends you hang out with, how im
portan t is the following? Most heavily loading
items are attending class regularly, studying, get
ting good grades, and continuing education past
high school.
Second factor (eigenvalue=2.00) derived from a
factor analysis of eleven items following the qu es
tion, Am ong the friends you han g out with, how
important is the following? Most heavily loading
items
are having
steady boy friend/ girlfriend and
being willing to party and get wild.
Proportion of alters in ego's nominated network
who nominated at least one other alter in this
network as a mem ber of their own friendship net
wo rks. Based on alters wh o respond ed to our
questionnaire.
Quotient of numerator and denominator, where
denom inator is the ratio of ego's observed to pos
sible nominations received from the alters in the
egocentric network. Denom inator
is
the average
of
the comparable ratios, calculated for each of ego's
nominated friends (alters). In both numerator and
denominator, possible nominations received are
limited to those from people in ego's network w ho
responded to our qu estionnaire.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 148/375
131
II
Distributions of Variables.
All distributions are or
the
sample used in
Table
except for No n-Spo rt Extracurricular
Participation and corresponding riends mean (samplefrom Table
1),
School Troubleand
corresponding friends mean (sam ple from Table 2), and E ducational Expectations and
corresponding friends mean (sample from
Table
4).
Dependent Variables
Non-Sport Ext racurr icu lar
Part icipat ion
School Trouble
Job Kn ow ledge S core
Educat ional Expectat ions
Independent Variables
Female
Paren t ' s Educat ion
Professional I
Professional II
M a n a g e r / A d m i n i s t r a t o r
Laborer
Other Blue Collar
Forest View
River
Midd leb rook
Roosevel t
Magne t
G P A
Senior
College Prep Track
Voc /Tech Track
Me an of Fr iend s ' of No n-Sports
Mean of Friends ' School Trouble
Mean of Friends ' Job
Knowledge Scores
Mean of Fr ien ds ' Educat ional
Expectat ions
Aca dem ic Fac tor
Social Factor
Cohesion
Central i ty
Mean= 1.78
Mean= 3.43
Mean= 1.38
Mean= 7.22
1 62.5%
Mean= 4.10
1 11.2%
1 7.6%
1 13.6%
1 10.5%
1 31.9%
1 6.9%
1 18.1%
1 23.9%
1 12.7%
1 21.6%
Mean= 3.15
1 39.7%
1 51.8%
1 6.2%
Mean= 1.59
Mean= 3.47
Mean= 1.22
Mean= 7.26
Mean= 0.07
Mean=-0.10
Mean= 0.31
Mean= 1.17
Std. Dev.= 1.51
Std. Dev.= 2.92
Std. Dev.= 1.53
Std. Dev.= 1.81
0 37.5%
Std. Dev.= 1.87
0 88.8%
0 92.4%
0 86.4%
0 89.5%
0 68 .1%
0 93 .1%
0 81.9%
0 76 .1%
0 87.3%
0 78.4%
Std. Dev.= 0.80
0 60.3%
0 48.2%
0 93.8%
Std. Dev.= 1.24
Std. Dev.= 2.37
Std. Dev.= 1.19
Std. Dev.= 1.33
Std. Dev.= 0.91
Std. Dev.= 0.97
Std. Dev.= 0.19
Std. Dev.= 0.83
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 149/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 150/375
School Choice and
Community Segregation:
Findings from Scotland
/. Douglas Willms
University of New runswick
and
University
of dinburgh
Many educators, parents and politicians are optimistic that schemes
designed to increase choice in schooling will bring ab out an imp rove ment
in Am erica's
schools.
The early pro ponents of increased choice argued that
the use of free market mechanisms would increase competition among
schools and make schools more accountable to parents (Freidman 1962;
Hirsch ma n 1970). They proposed voucher schemes whereby p arents w ould
receive a voucher that they cou ld app ly to tuition costs at either a public or
priv ate school (e.g., see Coons
Sugarman 1978), or alternatively, tuition
tax credits, which w ould allow paren ts to cover some of the costs of p rivate
schooling through tax deductions.
The State
of Minnesota has allow ed such
ded uctions for som e time. Recently, mechanism s to increase choice within
the public sector have been introduced in a num ber of states (see Clun e
W itte 1990). These have taken a variety of forms such as mini-schoo ls
within schools, open enrollment policies, magnet schools, and controlled
choice pro gra m s (Rassell & Rothstein 1993). Many of the mechan isms are
intended
to
increase
the
diversity of school offerings and improve
the
match
betwee n school prog ram s and pa ren ts' and stude nts' preferences. Because
increased choice me ans schools m ust com pete for stud ents, its prop onen ts
hope that school and classroom practices will improve and that higher
academic stan dards w ill be achieved.
There are few critics of increased p arenta l choice.
It was
endorsed by past
Presidents Reagan and Bush, and by President Clinton. It has received
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 151/375
134
widespread support f rom bus iness -or iented conserva t ives , l ibera l pol icy
scholars , the Afr ican-American communi ty , and the Cathol ic Church
(D ou gh erty Sostre 1992). I ts su pp or ters arg ue that increased choice wil l
g ive poor and d isadvantaged famil ies opportuni t ies tha t have long been
available to the white middle-class . But some educationis ts are concerned
that increased choice m ay further e l i tism in schools : families w ith grea ter
social an d c ultural cap ital wil l hav e greater abil ity to exercise choice, w hich
will lead to greater inequalities along racial, ethnic, and social class lines
(Levin 1980). De spite con siderab le discussion am on gst resear che rs , there
has been l i t t le empirical evidence on the effects of choice programs in the
U.S.
Some of the f irs t analyses of data from the National Educational
Lo ngi tu dina l Study sugges t tha t choice pro gra m s are no tbr ing ing ab ou t the
int en de d g ains in acad em ic achievem ent (Schiller 1993); ho w ev er, the
effects are difficult to discern because there is such wide variation in types
of choice programs in the U.S. (Plank
et al
1993).
In the U.K. legis lation tha t enables pa ren ts to choo se schools w ithin the
pub lic sector ha s bee n in place for over a decad e. The 1980 Ed uca tion Act
and the 1981 Education (Scotland) Act gave parents the s tatutory r ight to
request places in schools outs ide their designated at tendance areas . The
legis lation also required local edu cation auth ori t ies (LEAs, cal led EA s in
Scotland) to publish brochures for each school that reported the school 's
exa m inatio n resu lts . Th e bro chu res w ere also to describe policies concern
ing ho m ew ork , uni forms , and school d isc ip l ine , and to inc lude informa t ion
abou t the curr iculum . The major ity of p lac ing reques ts ha ve been m ad e on
behalf of chi ldren enter ing pr imary school , or t ransferr ing to secondary
school. The Acts require education authori t ies to take these requests into
acco unt, and only und er certain circum stances can they be rejected. D urin g
the f irs t few years after the Acts were passed, nearly al l requests were
gra nte d. The B rit ish experience is relevan t to U.S. policy beca use in m an y
respects the reform embodies the ideals of a free-market approach s tr ived
for by the Am erican pro po ne nts of choice (e.g., see C hu bb M oe 1990).
Much of the empirical research on school choice has been conducted in
the U.K., ow ing heav ily to su pp or t from the Econom ic an d Social Research
Council . Scotland in part icular has had an act ive research program, led by
Michael Adler and his colleagues (Adler
1993;
Adler Raab
1988;
A d l e r
t
al 1989;
Petch 1986) an d furth ered by th e Ce ntre for Educ ation al So ciology
(Echols
et al
1990; W illm s Ech ols 1992).
The Adle r
et
al
(1989) s tud y, based on intervie ws w ith ov er 600 pa ren ts ,
found tha t paren ts w h o exercised choice w ere m ot iva ted m ore by a des i re
to avoid the school in their a t tend anc e area than to f ind the op tim al school
for their child. The reasons parents gave for choosing a part icular school
ten de d to be associated w ith social factors such as discipl inary cl imate or a
school ' s genera l reputa t ion , or wi th prac t ica l cons idera t ion su ch as proxim-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 152/375
235
ity. Few pa rents em phasized educational considerations such as teaching
m ethods or examination results.
The Echols
et h
(1990) study was based on a large nationally- represen
tative sam ple of the cohort of pup ils that en tered secondary school in 1982,
the first year that the legislation became
operational.
They found that better
educated parents and those with higher levels of social class were more
likely to exercise choice. Moreover, the schools they chose tend ed to have
higher social class intakes and were m ore likely to be schools that had been
founded before the turn of the century. Many of these older schools had
formerly been selective grammar schools, and still include the term acad
em y in their nam e. Willms and Echols (1992) later exam ined survey data
describing the cohort that entered secondary school in
1984,
and w ere able
to match some of the pupils ' responses with parent data from the Adler et
l sam ple. They found that parents w ho exercised choice chose schools with
higher mean socioeconomic status than that of their assigned school.
However, the chosen schools only marginally benefitted their children in
terms of academ ic attainm ent and were not particularly effective or ineffec
tive when compared with schools of similar SES intake.
Prior to the mid-sixties, Scotland operated a tri-partite system of second
ary mod ern, com prehensive, and gramm ar schools. Selection into second
ary schools was based primarily on teachers' accounts of pupil ability.
Du ring the late sixties and seventies, secondary schooling was reorganized
along comprehensive lines, and by 1980, when legislation on parental
choice was introd uce d, over 95% of Scottish pu pils attend ed their
neighbourhood comprehensive school. Reorganisation had the effect of
reducing segregation an d narrowing the gap in attainment between w ork
ing- and middle-class pupils (McPherson Willms 1987). Parental choice
may be reversing this trend. Both the Adler group and the
CES
researchers
expressed concerns that parental choice was causing an increase in segrega
tion along social class
lines,
and that this in turn would increase the gap in
educational attainment between children w ith middle- and working-class
backgrounds. However, these effects could not be directly observed with
cross-sectional da ta.
This study attem pts to further the work by em ploying longitudinal data
on schools and communities to address four questions concerning the
effects of the parental choice legislation over the period 1982-1991. First,
what is the incidence of parental choice in Scotland? Has de m and increased
as the reform has taken hold, and if so, has there been a decline in the
proportion of placing requests granted? Second, has the relationship be
tween choice behav iour and paren tal background remained stable over the
ten-year period? This question is asked with regard to choice within the
state sector, and the choice of independent schools. Third, have parents
tended to choose schools with higher social class intakes? Fourth, have
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 153/375
236
school ing sys tems in Scott ish comm unit ies becom e increasingly segregated
alon g social class lines, and if so, are increases in betw een-sch ool segre ga
tion related to the incidence of parental choice?
ata and M ethod s
The data come from two sources. One is a set of policy documents
pr ep are d b y The Scottish Office (1993,1992; Scottish Ed ucatio n D ep art m en t
1983,1985,1986,1987a, 1987b, 1989,1990) . These documents descr ibe the
incidence of choice with in the s tate sector from 1981 th ro ug h 1991. The
sec on d s ou rce is the Scottish You ng Pe op le's Surv ey s (SYPS) of 1985 ,1987,
1989,
an d 1991. The target p op ula tion s for the coho rt elem ent of the se
surv ey s w ere al l pu pils w ho s tar ted their fourth year of seco nda ry schoo l in
a Scott ish secondary school, ei ther public or pr ivate, in the preceding
academic year . Thus, the major i ty of the pupils covered by these surveys
ente red s eco nda ry school in 1980,1982,1984 and 1986. Pup ils w ere ad m in
is tered a questionnaire covering var ious topics regarding their school
exper ience , family background, educat ional a t ta inment , and pos t- school
destinations. In the last three of these surveys, half of the questionnaires
inc luded ques t ions pe r ta in ing to whe the r the pupi l s had a t t ended a
no n-d esig na ted school as a result of pa ren tal choice. Ach ieved s am ple s izes
for the four surveys were as follows: 1985, N = 6 5 0 1 ; 1987, N=6360; 1989,
N = 5 5 8 1 ; and 1991, N= 4450. Furth er details of the survey s are repo rte d by
L a m b ,
Burnhill and Tomes (1988).
Many of the analyses in this s tudy are s imply descr iptive, and are
por t rayed through s imple l ine graphs and his tograms. But the ques t ions
con cern ing the effects of choice across co m m un ities are mo re comp lex. O ne
of the central features of this analysis is that communities , schools , and
pupils are treated as the units of analysis in a multi level model. Because
near ly a l l of the movement across school ca tchment boundar ies or to the
priv ate sector occurs in comm un ities w ith tw o or m ore sch ools , the effect of
choice on segregation can be best observed at the community level . The
analysis est imates the extent of among-school segregation in each of 54
Scot tish com mu nit ies us ing three separa te segregat ion indices , descr ibed
belo w . Th e m ulti level mo de l is then u sed to test w he the r levels of se grega
t ion hav e been increasing in these comm un ities , an d if so, w he the r increases
are associated with the incidence of choice. Because the extent of choice
varies among the f if ty communities , and over t ime, we have a powerful
design for test ing whether choice is related to these indicators , both
cross-sectionally a nd long itudina lly. In this respect , the s tud y is un iq ue in
that i t exam ines dif ferences amo ng mult i - school com m un it ies in a m ul
t i level f ramew ork, w ith t ime as one level of the ana lysis .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 154/375
137
Definition of Variables
Social Class
Th e occupa tions of pu pi ls ' fathers we re classif ied in to the
seven categories of the Registrar General 's social-class categories: Profes
sional, Intermediate, Skil led Non-Manual, Skil led Manual, Par t ly Skil led,
Unskil led, and No Occupation or Unclassif ied (Off ice of Population Cen
suses an d S urv ey s 1970) . This scale ha s be en used extensively in British
educational and sociological research.
Parental Education
This m easu re com pr ises informat ion on the nu m be r
of yea rs of formal scho oling com pleted by the pup ils ' m oth ers an d fathers .
I t ha s three levels : bo th pa ren ts schooled to age 15 years or less; on e or b oth
pare nts schooled to age 16 years , bu t nei ther to 17 years ; one or bo th p aren ts
schooled to 17 years or m ore . W here only one paren t w as repor ted by the
pupil , the classif ication was based on that parent. The Scott ish schooling
system has always allowed pupils to achieve cer t if ication in the terminal
schoo l ex am inat ion s at arou nd age 17; therefore, the hig hes t level of this
var iab le can be cons idere d a prox y for the at tain m en t of this cer tif ication.
Socioeconomic Status A statis t ical composite of fathers ' occupation,
mothers ' educat ion, and number of s ib l ings was der ived through a pr inc i
pal com pon ents an alys is of the pu pi l s ' repo r ts on these measu res for 1991
SYPS data. The sam e scaling was then ap plied to calculate an individ ual- level
SES score for each pu pi l in each coho rt . Pu pi ls ' individ ual- lev el SES scores
were aggregated to the school and community levels to es t imate mean
school and community SES.
Community
The sam e def ini tion of com m unity em plo yed by M cPherson
& Willms (1987) w as u sed to classify scho ols into com m un ities across th e
four cohorts . They def ined community in three s tages:
irst
all schools,
including private
schools,
were allocated to places
as
defined
by the R egistrar G eneral, Scotland (RGS 1967). Second, all places served by
more than one school were then identified. T hird, if the schools in that place
served all pup ils in that place, and only those
pupils,
the place was defined as
a multiple-school community. In setting this definition, only data on school
location and catchment w ere used, and not details of individual p up ils' hom e
addresses. W here a school in one place served pupils living in another p lace,
the places were concatenated until a set of places was identified such that it
was the m inimum set within w hich the schools served all pup ils in that set,
and only those pupils. This set was also then defined as a multiple school
community, (p. 517)
W ith th is def ini t ion of com m unity , v i r tua l ly al l pu pi ls in the com m un ity
had no effective choice other than to at tend a school in that community.
Very few pupils in s ingle-school communities exercised choice. For ex
am ple, in the 1990 91 data, 20 of th epu pilsa tten ded schools in s ingle-school
com mu nit ies , bu t only 29 (6 .8 ) of the 424 pu pi ls w ho w ere a t ten din g
non-local schools were in s ingle-school communities .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 155/375
138
Segregation I em plo y three me asu res of segregation. Tw o of these are
based on the d i s t r ibu t ion of pup i l s wi th pa ren ts f rom middle - and
wo rking-c lass bac kg roun ds . In these analyses m idd le c lass compr ises the
top two ca tegories—profess ional an d interm edia te occupa t ions . O ne index
is the dissimilar i ty index, which indicates the proportion of pupils with
middle c lass backgrounds tha t would have to change schools in order to
achieve an even dis tr ibution across schools of pupils f rom middle- and
working-c lass backgrounds . The second index, the isola t ion index, indi
cates the extent to w hich m id dl e class pu pils are expos ed only to each o ther ,
and thus isola ted from pu pi ls wi th w orkin g c lass bac kg rou nd s . This index
differs f rom the dissimilar i ty index in that a minority group could be
un ev en ly dis tr i bu ted across schools , bu t if i t w ere a relat ively sma ll gro up ,
i t would not necessar i ly be isolated from members of the major i ty group.
The third index , the correlat ion ratio, indicates the prop ortio n of var ia nce in
ind iv idua l l evel S S that is am on g schools . This index is useful h ere in tha t
i t facil itates com pa rison s with ea r l ier w ork , an d w ith SES segre gatio n in
othe r cou ntr ies . All three indices are s tan da rdiz ed such that they vary from
zero to one. Massey and Denton (1987) descr ibe these indices in greater
detail .
esults
The Extent of Parental Choice
The incidence of placing requests for Scotland rose from just over 1% in
19 80 /81 to ab ou t 2.5% by 198 5/86, an d rose only s l ightly thereaf ter. By
1 9 9 0 / 9 1 ,3 %
of the secon dary pup i l po pu la t ion h ad plac ing requ es ts m ad e
on the ir
behalf
However , the percentages di f fer markedly among Educa
tion Authorit ies (EAs) . As one would expect, the incidence tends to be
high er in EAs serving urb an areas . Glasgow has had the larges t prop or t io n
of placing requests over the last four years , with 5% making requests in
1990/91 . Renfrew, which inc ludes the conu rbat ion com pr is ing Pais ley, an d
Lothian, which inc ludes Edin burg h, a lso had con s iderably high er pe rcent
ages than the nat ional average . Grampian, which inc ludes Aberdeen, i s
so m ew ha t of an exception: it ha d a 4% inciden ce in 1986/87, bu t by 199 0/
91 the level had dropped to about 2%.
Th e majori ty of requests at the seco nda ry level w ere m ad e on behalf of
pu pil s wh o w ere enter ing th eir firs t year of seco nda ry schoo l (referred to as
SI). Figure 7.1 sh ow s the nu m be r of requests for entry at SI , exp ressed as a
perc enta ge of the SI pu pi l po pu latio n. The incidence for Scotland rose from
8% in 19S3/84
1
to over 10% in 1987/88, and remained s teady at levels
be twe en 10% and 11% to 1990/91 . The provision al est im ate for 199 1/92 is
11.8% (The S cottish Office
1993).
Thu s, by the beg inn ing of the 1990's, m ore
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 156/375
P
e
c
n
a
g
o
P
P
a
o
r
o
O
C
O
N
>
O
0
1
—
I
—
I
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 157/375
14
tha n 10% of the seco nd ary school pu pil s were at tend ing a school other than
the school in their designated catchment area.
T he 10% figure does not ade qu atel y sho w th e impact of the legis lat ion for
m an y of the author i t ies . In Glasgow , Lothian, an d Renfrew, a bo ut 5 to 18%
of the pu pil s have ha d requ ests m ad e on their behalf du rin g the last half of
the dec ad e. Requ ests in Tay side wer e also high er than the na tion al f igures
since the introd uctio n of the legis lat ion. M oreov er , Adle r t
al
(1989) found
that wi thin Lothian and Tays ide , there were some c i ty areas where the
propor t ion was more than 50%.
In Scotland, an EA could refuse to gra nt
a
reques t if i t dee m ed com pliance
w ou ld be ser iously detr im enta l to the ord er or discipline of the schoo l or
to the edu catio nal well-being of the pu pils there , or if com plian ce m ea nt
tha t a school m u st app oin t anoth er teacher , or s ignificantly e xten d or al ter
the school. [See Section 28A(3)(a) of the 1980 Education (Scotland) Act.]
W ha t dis t in gu ishe s Scott ish legis lat ion from the Act cove ring Eng land an d
W ales was th at an EA cou ld only refer to con dit ion s at the receiving scho ol,
whereas south of the border an LEA could deny a reques t i f compliance
would affect the provision of efficient education or the efficient use of
resou rces (Ad ler 1990). D em an d-sid e prov ision in Scotland wa s init ial ly
stronger than that of England and Wales also because of the appeal
pr oc ed ur e. In both cases a pa ren t could app eal a refusal to a com m ittee; in
England and Wales the appeal committee 's decis ion was f inal , but in
Scotland a pa ren t cou ld ap pe al a refusal f rom the com mitte e to the cou rts .
In Sco tland, wh en an ap pea l is up he ld, the EA m us t review the cases of al l
parents who have s imilar c i rcumstances , regardless of whether they have
had an appeal refused (Adler 1990).
In the first two years following the legislation, nearly all requests in
Scot land w ere granted . Bu tby 1983, som e Scott ish EAs bega n ca pp ing the
enro lme nts of po pu lar schools. Between 1982 /83 and 1990/91 the nat ion al
rate of refusals ha s cl imbed from aro un d 3% to nea rly 14%. Glasgow, which
has ha d the highes t percentage of reques ts , a lso ha d the highes t perce ntag e
of refusals—17.4%. L othian and D un ba rton also had refusal rates abov e the
nat ional average .
Social Class and Educational Background of Choosers
Table 7.1 sho w s the social class and ed uca tiona l ba ck gro un ds of pa re nt s
w ho se children atten ded a pr iv ate school or a s tate-funded school oth er
tha n the des ign ated school in their at ten dan ce area. Ov er the pe r iod cov ered
by the SYPS data, there was a s teady increase in the proportion of pupils
atte nd ing pr iv ate schools or non-local s tate schools , an d the dis tr ib utio n of
the typ e of pup il mak ing choices cha ng ed. In 1985, there w ere s trong ,
statistically significant relationships between choice of
a
pr ivate school an d
fath ers ' social class (chi-square= 378.3, df=6, p < .01) an d p ar en tal ed uc atio n
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 158/375
TABLE 7.1 Percentag e of Studen ts in Pr iva te Schools and Chosen Sta te-Funded Schools ,
Social Class of Father
Professional Oc cupat ions
Intermedia te Occupat ions
Ski l led Non-Manual Occupat ions
Ski l led Manual Occupat ions
Partly Skilled Occupations
Unski l led Occupat ions
No Occupation or Unclassified
Chi-Square' (df=6)
Level of Parents ' E ducat ion
On e or both to 17 Years or m ore
One or both to 16 Years only
Both to 15 Years or Less
Chi-Square (df=2)
All Pupils
and by Fathers '
1985
Pr iva te
School (n)
18.2 (279)
9.1 (1232)
3.9 (386)
0.6 (2094)
0.3 (824)
0.3 (240)
2.1 (914)
378.3**
14.9 (965)
2.9 (1661)
0.6 (3583)
468.2**
3.4 (6209)
Occupat ion and Parents ' Educat ion
1987
Non-Local
Pr iva te Sta te
School School (n)
18.7 10.6 (173)
10.8 6.2 (632)
4.1 7.9 (200)
0.7 5.7 (1000)
0.6 4.1 (327)
0.0 4.7 (95)
1.5 5.6 (489)
208.7**
16.3*
16.9 8.1 (510)
2.7 6.0 (885)
1.0 5.3 (1575)
246.5**
11.0**
4.2 6.0 (2970)
'The chi -square tes ts d isplayed in the pr iva te school colum ns are tes ts for in
pr iva te /s ta te- funde d by socia l c lass .
the chi-square tests displayed in the
schools .
p< .05,
p< .01.
1989
Non-Local
Pr iva te
School
28.2
11.6
3.6
0.3
0.5
0.0
1.7
290.7**
16.6
3.7
0.9
186 . 9
5.1
idependen
State
School
n
8.3
8.5
9.1
9.0
5.9
8.8
1.6
8.9
9.2
8.2
2.3
8.7
(n)
(157)
(564)
(176)
(739)
(300)
(93)
(509)
(512)
(892)
(1138)
(2542)
ce of th e 2 by
and the 2 by 3 cont ingency table of pr iva te /s ta te - fund ed by
by Year,
1991
Non-Local
Pr iva te Sta te
School School (n)
21.6 10.1
10.9 9.9
3.6 9.6
0.9 9.8
0.8 11.6
0.0 9.6
3.6 12.3
286.2**
5.1
15.3 9.1
2.8 11.7
1.3 10.1
270.4** 2.6
5.3 10.5
(262)
(1036)
(289)
(1246)
(474)
(147)
(688)
(1006)
(1605)
(1532)
(4143)
' contingency table of
•
level of education. Similarly,
non- local s ta te school colu mn a pply to compar ison s of loca l to non- local s ta te- fund ed
ife,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 159/375
42
(chi-square=468.2, df=2, p < .01). The majority of pupils who attended
priva te schools were of middle class backgrounds, or had at least one paren t
schooled to 17 years or
more.
These relationships for choice into the p rivate
sector did not change significantly between 1985 and 1987. Recall that the
1987 data describe the first cohort that entered secondary school un de r the
pa ren t choice legislation. For that cohort the relationships between choice
of
a
non-local state school and fathers' social
class
and parental educational
were sim ilar to the distributions for choice of private
schools.
3
The relation
ships were not as strong (for fathers' social class,
chi-square=16.3,
df=6, p <
.05;
for p aren tal education, chi-square=11.0, df=2, p < .01), bu t there was
clearly greater propensity to exercise choice among higher social class and
better educated parents.
By 1989 there were tw o noteworthy changes. First, the pattern of choice
within the state sector changed, such that there were roughly equal prop or
tions choosing from each social class group, and for groups with similar
levels of education. Second, there was a steady increase in choice of the
pr ivate sector, rising from 3.4 in 1985 to 5.3 in
1991.
The relationship
between choice of the private sector and parental background remained
strong and statistically significant (for fathers' socialclass, chi-square=290.7,
df=6, p <
.01;
for parental education, chi-square=186.9, df=2, p <
.01).
The
increase in choice of the private sector did not stem from increased prop or
tions of working class pupils; rather, it stemmed from two factors. First,
there w ere rising levels of social class between
1985
and
1991;
for exam ple,
the proportion of pupils from professional occupations rose from 4.5 to
6.3 ,
and the proportion from intermediate occupations rose from 20 to
25 .
Second, despite the change in the social class distribution, the pro po r
tion of pupils from professional and intermediate backgrounds choosing
the private sector did not decrease; indeed, it increased slightly from 18.2
to 21.6 for professional occupations, and from
9.1
to 10.9 for interm e
diate occupations.
Socioeconomic Status of
hosen
Schools
Previous analyses of the Scottish data suggested that parents chose
schools with disproportionately high levels of mean pup il SES (Echols et
al
1990; W illms Echols 1992). Willms Echols (1992) com pared the m ean
SES
of the pupi ls' designated school with the chosen school they attended,
for a sample of
195
pup ils that were part of the Adler et al (1989) sam ple.
They found that on average the chosen schools had a mean SES that w as 0.25
of
a
standa rd deviation higher than their designated
schools.
The size of the
difference varied among the three regions studied . The SYPS da ta do not
identify the pu pil s' designated
schools;
thus, in this study I cannot estimate
differences in mean SES between designated and chosen schools. Instead,
I estimate the discrepancy between the SES of the chosen school and the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 160/375
43
TABLE 7.2: Differences between the Mean S S of Chosen Schools
and the Mean S S of the Co mm unity
Mean Standard
Difference Error N
1987
School Mean SES minus Mean S S of all Schools
in the Commu nity including Private) .041 .035 172
School Mean SES m inus Mean S S of State-Funded
Schools in the Community .127** .339 172
1989
School Mean S S minus M ean S S of all Schools
in the Commu nity including Private) .007 .290 207
School M ean S S minus M ean S S of S tate-Funded
Schools in the C om munity .101** .282 207
1991
School M ean S S m inus Mean S S of all Schools
in the Comm unity including Private) .032 .215 395
School Mean S S minus Mean S S of S tate-Funded
Schools in the Community .107** .215 428
m ea n SES of the co m m un ity . Mea n SES of the co m m un ity is calculated in
tw o way s : one tha t inc ludes bo th s ta te- funded and pr iv a te schools, an d one
that includes only the s tate-funded schools . The results are reported in
Table 7.2.
The results indicate that the m ea n S S of the chose n s tate-fund ed schools
do no t differ s ignif icantly f rom the m ea n SES of the com m un ities in w hic h
they are located, w h en pr iv ate schools are inclu ded . O bse rved differences
are small—less tha n .05 of a s ta nd ard dev iation — and stat is t ically insignif i
cant. H ow ev er , the m ea n SES of the chosen s tate-fun ded schools are abo ut
.10 to .13 of a s ta nd ard dev iation hig her than the m ea n SES of the co m m u
ni ty , w he n only s ta te- funded schools are cons id ered. There does no t a pp ear
to be a ny cha ng e in this effect ove r the per iod cov ered b y the SYPS dat a. W e
can conclude tha t parents choos ing within the s ta te sec tor d ispropor t ion
ately cho se schoo ls w ith high er m ea n SES tha n othe r s tate-sector schools .
4
hanges in omm unity Segregation
The two pre vio us an alyses show ed tha t there w as a smal l increase in the
prop or t io n of pare nts choos ing pr iva te schools , and th a t pare nts w ho chose
private schools tended to be of higher educational and social class back
ground. The analyses a lso revealed tha t those choos ing s ta te- funded
non- local schools dispropor t ionate ly chose high SES schools . Taken to
gether , these two f indings suggest that parental choice may be causing an
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 161/375
44
TABLE 7.3 Parameter Estimates for HLM Regression Models Explaining
Levels and A nnual C hanges in Com munity Segregation Indices
Dissimilarity Index
Average Level of Segregation
Annual Increase in Segregation
Propo rtion Exercising Choice
Isolation Index
Average Level of Segregation
Annual Increase in Segregation
Propo rtion Exercising Choice
SES Segregation Index
Average Level of Segregation
An nua l Increase in Segregation
Propo rtion Exercising Choice
Model 1
Estimate
.2467**
.0085
.3225**
.0089**
.0971**
.0070*
(SE)
.0233)
.0047)
.0159)
.0029)
.0164)
.0030)
Model 2
Estimate
.2453**
.0058
.0909*
.3225**
.0044
.0489
.0971**
.0044
.0283
(SE)
.0233)
.0064)
.0433)
.0159)
.0045)
.0358)
.0164)
.0046)
.0367)
*p<.05
**p<.01
increase in the extent of among-school SES segregation in local communi
t ies.
Tab le 7.3 pr ov ide s estim ates of the effects.
M o d e l sho w s estim ates of the ave rage intercepts an d avera ge s lop es for
the w ithin -co m m un ity regression s of the segreg ation ind ices on year . Year
w a s co d ed 0 for 1985, 2 for 1987, 4 for 1989 and 6 for 19 91, su ch th at th e
intercept represents tha t average value of the index am on g the 54 co m m u
nities for 1985. The estim ate for the dissim ilarity ind ex is 0.2482, wh ich
indica tes tha t approx imate ly one-qu ar ter of the midd le c lass pu pi ls w ou ld
ha ve h ad to chang e schools to achieve an even d is tr ibution of pu pi ls ac ross
a l l schools . The es t imated ann ual increase was 0.0085 pe r yea r (statistically
significant at p<.10, but not at p<.05). Using this figure, the estimated
ch an ge in the dis sim ilarity ind ex for the per io d 1985 to 1991 is therefo re (6
x 0.85=) 0.05; that is, an increas e from 0.25 to 0.30. Th us, to ach ieve an eve n
dis t r ibut io n of m idd le an d work ing class pu pi ls , an add i t iona l 5 of a ll
pu pi ls w ou ld n eed to chang e schools . To place the values of these segrega
tion indices in context for the American reader, the Dissimilarity Index for
Blacks in 1988 in San F rancisco w as .235; in Los An ge les it wa s .530 (see
Rumberger & Willms 1989) .
The estimate of the average isolation index across the 54 communities
w as 0.3250, an d th e av era ge an nu al increa se w as 0.0089 (statistically signifi
cant , p<.01). This indic a tes tha t appro xim ate ly three -e ig hth s of the
m iddle-c lass pup ils were expos ed only to each other , an d th us isolated from
working-c lass
p u p i l s .
The significant an nu al increase indicates that over the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 162/375
245
FIGURE 7 Average of Segregation Indices for Scottish Comm unities
©• 5 Cities: Isolation *- Others: Isolation
© 5 Cities: Dissimilarity * Others: Dissimilarity
e 5 cit ie s: Eta for
SES
— * Others: Eta for SES
pe riod 1985 to 1991 there w as a tend enc y for mid dle class pu pi ls to bec om e
more isola ted in schools tha t were predominant ly middle-c lass .
The las t m eas ure of segregat ion, wh ich is calcula ted us ing our c om posi te
m ea su re of SES, indica tes the prop ort ion of var ianc e in SES tha t is am on g
schools . The ave rage level of segreg ation o n this m ea su re w as .0971, with an
estimated annual increase of 0.0070 (statistically significant, p<.05).
The estimates of the segregation indices and their annual change var ied
across the
5
com mu nit ies . Some of th is var ia t ion
is
attr ibutable to sam pling
error ; with these da ta i t is not po ssible to achieve accurate est im ates of the
indices for the sm aller co m m un ities (see W illms Paters on 1995). H ow
ever , it is possible to achieve reaso nably accu rate est imatio n of the indices
for the larger communit ies , and for the smal ler communit ies combined.
Figure 7.2 displays the changes in the average of each of the segregation
indices for Scotland's f ive largest ci t ies (Glasgow, Edinburgh, Pais ley,
Aberdeen, and Ehandee) and for the other 49 communities . For the f ive
largest
cities,
taken to gether , al l three indices increased o ver the pe r io d. T he
f igure show s tha t segregat ion increased for both large an d smal l co m m un i
ties alike, an d tha t the bigg est increase wa s for the isolation of m id d le class
pupils in the five large cities.
Fig ure 7.3 di sp lay s the increase s in the isolatio n ind ex for eac h of th e five
cities.
Isolation increased between 1985 and 1991 in all of these cities,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 163/375
246
FIGURE 7.3 Isolation of M iddle Class Pupils in 5 Scottish Com mun ities
85 86 87 88 89 90 91
Year
without except ion. However , the change in Edinburgh, which had the
high es t level in 1985, w as smal l . A berd een 's index a lso cha ng ed only
s l ight ly . The changes in Glasgow, Pais ley, and Dundee, however , were
substantial . The change in Glasgow, Scotland's largest ci ty, is par t icular ly
al ar m in g: it ros e from .392 in 1985 to .637 in 1991. The
99
level of isolation
is higher than the extent that Blacks were isolated in any of the six major
scho ol distric ts in California in 1988.
Changes in the SES segregat ion index provide some indica t ion of the
im por ta nce of the changes in Glasgow . McPherson & Wil lms (1987) docu
mented a small but s teady decline in the segregation index for Glasgow
be tw ee n 1977 an d 1985, usin g da ta f rom the schoo l- leaver elem ent of the
SYPS. Their est ima tes of
SES
seg rega tion for Glas go w w ere 0.37 (1977), 0.35
(1981) an d 0.33 (1985). The estim ates calcu lated in this stu d y for G las go w ,
using the cohort element of the SYPS were: 0.39 (1985), 0.44 (1987), 0.44
(1989),
and 0.56 (1991).
The second m ode l in Table 7 .3 exam ines w heth er the ch ange s in the
segre gatio n indices ov er the per io d for each co m m un ity w ere related to the
pro po rtio ns of pupi ls exercising choice. If choice w ere cau sing a n increase
in segregation, we would expect to observe larger annual increases in the
segregat ion indices in comm unit ies w here the re was a higher inc idence of
choice. For all three indices, the an nu al rate of increase is par t ial ly exp laine d
by the extent of choice in the community. For the dissimilar i ty index, the
incidence of choice accounted for nearly all of the annual increase in
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 164/375
47
segrega tion. For the other two indices, the incidence of choice accounted for
about one-half of the annual increase. The coefficient for the proportion
exercising choice
is
statistically significant for the dissimilarity index
(p<.05).
For the other tw o ind ices, the coefficients are in the expected direction, bu t
are not statistically significant. One limitation of this analysis is that the
power of statistical tests at the community-level depends mainly on the
number of communities, not the number of pupils, and with only 54
com munities, the po wer to detect significant effects is limited.
Conclusions
Legislation in the early 1980s that enabled parents to choose schools
outside of their designated attendance areas brought a type of open
enro llment system to British schools. This stud y examines some of the
effects of that reform on the prov ision of secondary schooling in Scotland,
The analysis employs documents from the Scottish Office and data from
four of the bienn ial Scottish Young Peoples ' Surveys. One survey covered
the cohort of pupils that transferred to secondary school one year before the
reform became operational, and three surveys covered the cohorts that
transferred to secondary school du ring the first, third and fifth years after
the choice plan w as in
place.
This stud y exam ines the extent that the reform
has taken ho ld and its effect on school segregation in local com munities.
The incidence of paren ts mak ing reques ts for non-local schools climbed
rapidly during the first five years of the reform, and then levelled off.
Approxim ately 10 to 12% of the pup ils transferring to secondary school
have placing requests made on their
behalf
As the incidence of requests
increased, the proportion of requests refused also increased. During the first
few years of the reform, only about 2 to 4 percent of the requests were
refused, but by
1990/91,
the refusal rate had reached nea rly 14%. In some
education authorities, it was considerably higher. Many of the refusals
stemm ed from the capp ing of pop ular schools that had reached maxi
mum capacity.
The findings suggest that during the first few years of the reform, the
right to choose was exercised mainly by parents from middle-class back
gro unds . But as the reform took hold , a larger proportion of paren ts from
working class backgrounds exercised choice. These changes were accompa
nied by an increase in the proportion of pupils opting to leave the public
sector for private schools. The analysis shows also that paren ts disp ropo r
tionately chose schools that served pupils with above-average levels of
socioeconomic status.
The extent of segregation of middle- and working-class pupils was
estimated for each year for the
5
Scottish com munities that had at least two
secondary schools. The analysis shows that between-school segregation
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 165/375
48
alo ng social class l ines w as increasin g substan tial ly du rin g the pe r iod w he n
the choice reform w as pro ceed ing. Differences a m on g three type s of segre
gation indices examined suggest that one of the main tendencies has been
for middle class pupils to increasingly become segregated in a small
number of schools wi thin each community .
The extent of segregation and i ts rate of increase var ied substantial ly
am on g com m un ities . Because of this var iat ion , an d becau se the incide nce of
choice var ied among communit ies , the analys is could tes t whether the
inciden ce of choice w as related to the an nu al increase in segregatio n, usin g
pu pi ls , schools , an d co m m un ities as levels of analysis in a m ulti leve l m od el.
Th e incidenc e of choice in a co m m un ity w as posit ively related to the an nu al
increase in segregat ion, and w hen e ntered in to the m od el expla ined abo ut
half of the annual increase in segregation. However , with only 54 commu
nit ies, the relat ionship between choice and increase in segregation was not
statistically significant.
This s tudy cannot s tate def init ively that choice legis lat ion
caused
an
increase in school segrega tion. Ap art f rom the po w er of the analy sis , there
are a n u m be r of threats to the internal validity of such a
cla im.
One p r ob l e m
is tha t the incidence of choice is greater in larger com m un ities , an d larger
com mu nit ies tend to hav e grea ter be tween -school segregat ion. Du r ing the
pe r iod s tud ied, i t i s poss ible tha t res ident ia l segregat ion increased, par t icu
larly in the cities. The socia l c lass s t ruc tu re change d du r ing tha t per iod , w i th
high er prop or t ion s of pare nts in midd le c lass occup at ions . Levels of p aren
tal education also differed across the four cohorts s tudied, with the most
recent cohor ts having higher levels . Thus there could have been severa l
forces other than increased parental choice that brought about the increase
in com m un ity segregat ion. Never theless , the f indings pro vid e s t rong evi
denc e tha t choice is not he lping to reduce betwee n-school se gregat ion in
Scot t ish communit ies . While some of the proponents of f ree-market ap
proa che s to educ at ion argu e tha t choice m echan isms wil l im pro ve op po r
tunit ie s for pu pi ls from po orer ba ck gro un ds , the Scott ish exp erience refutes
this argument . I f anything, pupi ls f rom working c lass backgrounds are
be ing isolated in schools w ith less favourable contex ts .
Although the American and Scott ish schooling systems are very differ
ent , Scotland's experience with choice has implications for dis tr icts imple
m en tin g choice pla ns and other reforms in the U.S. In m an y respe cts ,
Scotland pro vid es a natu ral expe rime nt o n the effects of choice. Legislat ion
on choice beca m e op eratio nal in Scotland in 1982, and at that t ime near ly al l
pupi ls were a t tending schools in the i r des ignated a t tendance areas , wi th
less tha n 4 at ten din g pr iv ate schools (Mc Pherson & W illms 1987). A lso,
there wa s con siderab le uniformity in school goals , curr icula an d pr inc iples
of org aniz ation (McP herson & Raab 1988). Because of this unifo rmity , the
effects of m ark et reforms are m or e easily discerned tha n in m an y schooling
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 166/375
49
sys tem s in the
U.S.
Also, beca use S cotlan d's s tart ing poi nt wa s close to fully
co m pre he nsi ve ( i.e ., non-select ive), an d beca use the choice legis lat ion w as
far- reaching, the Scot t ish exper ience m ay pro vid e an up pe r b o u n d on
w ha t m ight be expected of s imi lar re forms in the
U.S.
Of course the f indings
are of im po rtan ce in their ow n right , to Scott ish educ ato rs in par t icula r, a nd
to the sociology of education in general .
Th e two pr inc ipa l s trategies for reform ing U .S schools in the 1990s are to
increase paren ta l choice and to res t ru c tur e schools , m ainly by giving
par ents , teachers , an d pr inc ipa ls grea ter au ton om y. One of the lessons f rom
Sc otland is tha t schoo l refo rm ers ne ed to pa y atte ntio n to the effects of these
re fo rms on the en t i re s choo l com m uni ty . A nu m ber of s tud ies w or ld w ide
ha ve sh ow n th at the collect ive pro per t ies of a school ha ve an effect ove r a nd
abo ve th e effects of pu pi ls ' ind ivid ua l bac kg ro un ds (see W illms 1992, for a
review). Thus the performance of a school is contingent on the types of
p u p i l s a tt e n d i n g
it
W hen a pu p i l w i th an adv an tag ed backg rou nd t rans fe rs
from a low - to a high-S ES sch ool, the con tex tual effect is s t reng then ed for the
cho sen school an d we ak en ed for the school the child left. U nlike the m ark ets
for con sum er pro du cts w her e al l bu ye rs can shif t from o ne br an d to a noth er ,
it is not possible for all pupils to attend a school with a high social class
intak e. If choice an d gre ater auto no m y are to s tren gth en the h an d of schoo ls
tha t a re a l ready ad van tag ed , w i tho ut safeguards to ensu re equal i ty , w e wi l l
l ikely see w iden ing disparit ies betw een the adv antag ed a nd disad van taged .
Notes
This pape r wa s prep ared for the 994 An nual M eeting of the Am erican Educa
tional Research Association. I am grateful to the Nuffield Fou nda tion for funding
the project, 'Standards, Tests and Parental Choice' which supported work on the
paper, to the United K ingdom Economic and Social Research Council for its support
of the Designated Research Centre at the Centre for Educational Sociology and
to
the
Canadian Social Science and Humanities Research Council.
1.
Figures for 1981/82 and 1982/83 were not available.
2.
The SYPS estima tes of choice with in the state sector for Scotland are w ithin
abo ut one-half of one percent of the figures rep orted by The Scottish Office. These
differences may be attributable to error resulting from pup ils w ho answered not
sure to the SYPS qu estion about choice, or du e to sampling error.
3. This finding was rep orted earlier by Echols
t
at (1990) for the 1987 cohort.
4.
The discrepancy between the
SES
of the community and the
SES
of the chosen
school
probably underestimates
the
extent
to
which those choosing schools increased
the SES of their school. For the 1989 data, Willms and Echols (1992) estimated
differences b etween chosen school and designated schools for
three
authorities. They
reported differences of
0 30
for Lothian, 0.10 for Tayside, and
0 21
for Fife, which are
much larger than the discrepancies between comm unity SES and chosen school SES—
0.10,0.07, and 0.03 for these th ree authorities in that year.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 167/375
250
e f e r enc e s
Ad ler, M ichael . 1993. An A lternat ive A ppr oac h to Parental Cho ice. Pp . 183-98 in
Briefings for the National Com mission on Education.
London : H e inem ann .
A dler, M ichael . 1990. Rights as Tr um ps: The Case of Paren tal Choice of School in
Scot land . Education and the Law 2:67-72.
Adler, Michael , All ison Petch, and Jack Tweedie. 1989.
Parental hoice and Educa-
tional Policy. Edinb urgh: Edinb urgh Univers i ty Press .
A dle r, Michael an d Gillian M. Raa b. 1988. Exit, Voice and L oyalty: Th e Im pa ct of
Paren ta l Choice on Adm iss ions to Secondary Schools in Edin burgh and D und ee .
Journal of
Education
Policy 2:155-179.
Co ons , John E. and Stephen D . Suga rm an. 1978.
Education by Choice.
B erkeley, CA :
Un iversi ty of Cal ifornia Press.
Dougherty , Kevin
J.
and Lizabeth Sostre . 1992. M inerva an d the M arketpla ce: The
Sources of the M ove m ent for School Cho ice.
Educational Policy
6:160-179.
Echols, Frank H., A nd rew F. M cPh erson, and J . D ouglas Wil lm s. 1990. Pa renta l
Choice in Scot land.
Journal o f Educational Policy
5:207-222.
Friedman, Mil ton. 1962. Capitalism and Freedom. Chicago: Unive rsi ty of Ch icago
Press.
Hirschman, Albert 0 . 1970.
Exit, Voice and Loyalty: Responses to Decline in Firms
Orga nizations and States. Cambridge , MA: Harvard Univers i ty Press .
Lamb, Joanne M., Peter Burnhil l , and Hilary Tomes. 1988. SYPS 1987: Technical
Report
(Edinburgh: Edinburgh Universi ty , Centre for Educat ional Sociology).
Levin, H en ry M. 1980. Edu cat ional Vou chers and Social Pol icy. Pp . 103-32 in
are
and E ducation of Young Children in America edi ted by Ron Haskins and James J .
Gal lagher . No rw ood , NJ: Ablex .
M cPh erso n, A nd rew F., and J. Do uglas W il lms. 1987. Equ al isat ion and Im pro ve
ment: Some Effects of Comprehensive Reorganisat ion in Scot land.
Sociology
21:509-539.
Office of Populat ion Censuses and Surveys. 1970.
Classification of Occupations.
London : HMSO.
Petch, All ison. 1986. Paren tal Choice in E leme ntary S chool .
Research Papers in
Education
1:26 47.
Plank , Stephen , Kathry n S. Schiller, Barbara Schneider, a nd Jam es S. Cole m an. 1993.
Effects of Cho ice in Ed uca tion . Pp . 111-34 in School Choice: Examining the
Evidence edi ted b y E dith Rasell and R ichard R othstein. Econom ic Pol icy Ins t i tute:
W ashington , D. C.
Registrar G eneral , Scot land
RGS).1987. AnnualEstimatesofthePopulation.
Edinburgh ,
H M S O .
Rasell , Edith and Richard Rothstein (Eds.). (1993).
School Choice.
W ash ing ton , DC:
Econo mic Pol icy Inst i tute .
Schi ller, Ka thryn . 1993, Feb ruary. Schools of Choice and Studen t Ac hiev em ent .
Paper presented at the Sociology of Educat ion Associat ion Annual Meeting,
Asi lomar , CA.
Scottish Edu cat ion D epa rtm ent . 1983. Placing Requ ests in Edu cat ion A utho ri ty
Pr im ary an d Sec ond ary Schoo ls. (Statistical Bulletin N o . 9 /B6/1983) . Edinburgh :
Government Stat ist ical Service.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 168/375
151
Scottish Education Departm ent. 1985. Placing Requests in Education A utho rity
Schools. (Statistical Bulletin N o. 2/B6/19 85). Edinburgh: Government Statisti
cal Service.
Scottish Education Department. 1986. Placing Requests in Education Autho rity
Schools. (Statistical Bulletin No. 5/B6/1986). Edinburgh: Go vernm ent Statisti
cal Service.
Scottish Education D epartm ent. 1987a. Placing Requests in Education Au thority
Schools. (Statistical Bulletin N o. 2/B6/1987 ). Edinburgh: Governm ent Statisti
cal Service.
Scottish Education Department. 1987b. Placing Requests in Education Au thority
Schools. (Statistical Bulletin
No.
9/B6 /1987 . Edinburgh: Government Statistical
Service.
Scottish Education Departm ent. 1989. Placing Requests in Educa tion Auth ority
Schools. (Statistical Bulletin
N o.
/B6/1989 ). Edinburgh: Governm ent Statisti
cal Service.
Scottish Education Departme nt. 1990. Placing Requests in Educa tion A utho rity
Schools. (Statistical Bulletin No. 3/B6/1990). Edinbu rgh: Governm ent Statisti
cal Service.
Scottish Office. 1992. Placing Requests
in
Education Autho rity Schools. (Statistical
Bulletin N o. Edn/B6 /19 92 /13 ). Ed inburgh: Governm ent Statistical Service.
Scottish Office. 1993. Placing Requests in Education Au thority Schools. (Statistical
Bulletin N o. Edn/B 6/1993 /13 ). Edinburgh : Government Statistical Service.
Willms, J. Douglas. 1986. Social Class Segregation and its Relationship to Pupils'
Examination Results in Scotland.
merican Sociological Review
51:224-241.
Willms, J Douglas.
1992.
M onitoring School Performance: A Guide for
Educators
Lewes:
Falmer.
Willms, J Douglas and Frank H, Echols.
1992.
A lert and Inert Clients: The Scottish
Experience of Parental Choice.
Economics o f Education Review
11:339-350.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 169/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 170/375
Educational Processes
and School Reform
Maureen
T
Hallinan
University o Notre ame
Schools serve thre e ma in pub lic functions in society: poli t ical , econ om ic
an d social (Spring, 1991). The poli tical function invo lves trainin g s tu de nt s
to be goo d c i t izens , pre par ing s tud ents for poli t ica l leade rship , t ransm it t ing
poli t ical bel iefs and ideologies , and maintaining power and s tabil i ty. In a
democra t ic soc ie ty , schools a im to preserve f reedom and teach s tudents
about their poli t ical and civi l r ights and responsibil i t ies . In a total i tarian
society, schools can be instruments of oppression and poli t ical control .
The econ om ic function of schools is to con tribute to coun t ry s economic
gro w th a nd technologica l develo pm ent . The school i s seen as the t ra in ing
place for future workers ; schools prepare s tudents for part icipation in the
labor m ark et . T hr ou gh schoo ling, s tu de nts are sorted by abil ity an d in terest ,
and prepared for fu ture employment .
The social function of schools requires preparing s tudents to l ive in
communi ty , to reduce soc ia l tens ions and conf l ic ts and to work toward
social jus t ice. Scho ols teach s tu de nts m or al va lues an d social respo nsibil i ty
in order to insure soc ie ta l s tabi l i ty and communi ty development . Often
schools are expected to play a role in social improvement, by direct ly
deal ing wi th soc ia l problems such as drugs , a lcohol ism, c r ime, v io lence ,
unemployment , and t eenage p regnancy .
In addit ion to these public functions, schools serve private functions.
G oo dl ad (1984) identif ies th ese functions as vocational , social , intel lectual
an d pe rso na l . Vocational goals are to train s tu de nts for jobs, social goa ls are
to prepare students for life in society, intellectual goals are to inculcate
knowledge and develop s tudent abi l i t ies , and personal goals a re to fos ter
s tudent ta lents , responsibil i t ies and freedom of expression.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 171/375
254
The pub lic and pr iva te goals of edu catio n gen erally coincide . The s truc
tures and processes that occur in school tend to facilitate the attainment of
bo th com m una l and ind iv idua l
goals.
At
t imes ,
how ever , publ ic and pr iva te
goals conflict . A utho rit ies may differ f rom in div idu al ci t izens w ith respec t
to pr ior i t ies , s tan da rds , and m etho ds of a t ta ining goals . In teres ted gro up s
also may differ in the emphasis they place on the var ious public goals of
education. Demands for school reform often s tem from differences of
opinion among public off icials and between public authorit ies and pr ivate
cit izens about which school goals are being met and how schools are
succeeding or failing.
Pressure for school reform, as evidenced in the current reform move
m en t, generally occurs dur in g a per iod of s ignificant social cha ng e. The late
1980 s an d ear ly 1990 s typif ies su ch a per iod. D ram atic eco nom ic, social
and dem og raph ic uphe ava l charac terizes th is per iod, c rea t ing cons ide rable
social turmoil . During unsett led t imes, public opinion tends to shif t , pro
ducing changes in at t i tudes and revealing value conflicts . Some values
receive par t ic ular at tention an d are translate d into policies an d pro gr am s by
individuals and groups . Pressure then is exer ted on the schools to make
changes that are consis tent with these public shif ts in values.
Several recent reforms can be directly linked to efforts to inculcate
specif ic values in the school. For example, Democratic and Republican
poli t icians recently have disagreed sharply about rel igious and moral
values. As a result , school prayer has become an issue of public concern.
Similar ly, in the 1980 s , a grow ing interest in rem ainin g technolog ically
compet i t ive in a broadening wor ld economy has pressured schools to be
ins t rum ents of economic growth and dev e lopm ent by prov id ing s tuden ts
w ith the skil ls req uire d for a cha ng ing labor m arke t. Today, as a result of an
increased public concern about personal safety, many school reform pro
grams a im to reduce drug and a lcohol abuse , h igh school drop-out ra tes ,
absenteeism and cr ime and violence in the schools .
Th e basis for the pre ssu re to reform schoo ls is the wid esp rea d v iew tha t
the public school is a major instrument of social change. This perspective
resu lts from the belief that schools can solve social pro ble m s. The form ation
and e xpa ns ion of the publ ic school sys tem in the ninetee nth an d twe nt ie th
centuries reflect the deeply rooted faith of Americans in the role of educa
t ion as an instrument of social mobili ty and societal advancement.
The reform movement also is sustained by poli t ical considerations.
Bow les and Gin tis (1976) ar gu e tha t the sour ce of social ills, such as racism ,
sexism, po verty , cr ime an d violence, l ies deep ly em be dd ed in the s tru cture s
of American society. I f these problems were addressed at their roots ,
considerable societal disruption and conflict would result. I t is far less
dis rup tiv e to rely on the s low proces s of socialization tha t occurs in schools
to sh ap e the at t i tu des an d beh avio rs of future ci t izens, than to add res s these
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 172/375
55
issues directly. At the same time, educators see a causal relationship
between deficiencies in the public school system and negative student
behavior (Mortimer, this volume), leading to greater immediacy to the
reform movement.
Since a call for school reform typically reflects value conflicts in society,
it
is
imperative to unde rstan d the mo tivation driving a particular reform in
order to estimate its potential to improv e student
learning.
For example, the
contem porary call to detrack schools, that is, to replace homog eneou s
ability group ing with heterogeneously grou ped classes,
is
really an effort to
rem ove social inequ ities in the schools. Advocates of detrack ing claim that
detracking decreases racial and socio-economic segregation, and provides
socially disadvantaged students with better opportunities to learn (see
debate between Hallinan and Oakes, 1994).
School Reform and Student earning
The paramount goal of schools, which should take precedence over all
others, is to increase student knowledge. Every other goal, public and
private, is an extension of this fundamental purpose of schooling. All
reform efforts ultimately should be directed toward improving student
learning. The danger of reform plans that stem from deep societal value
conflicts is that s trong attachm ent to a set of values and a sense of urgency
about institutionalizing them m ay distract from the prim ary imp ortance
of
the reform as a means to promote student learning. Many reform plans
seem to be designed not w ith student learning in mind, but rather w ith a
focus on other, often laudatory , goals. These reforms often are rem arkably
successful in achieving their purpose but do little to advance student
achievement.
Since stud ent learn ing should be the prim ary goal of every reform effort,
the success or failure of a particular reform must be determined on this
basis. To evaluate a reform, it is essential to identify the mo tivation of the
reformers and to determine whether their motivation is consistent with
and/or promotes student learning. The greater the growth in student
learning, the m ore successful
the
reform. Even if secondary public or p rivate
goals are met, the reform is an educational failure if it does not improve
student learning.
To evaluate the potential of a reform to increase student learning, the
links betw een the conditions of
the
reform and studen t outcomes should be
made explicit. It is essential to specify the connections between organ iza
tional, pedagogical or curricular change and the processes that govern
learning. U nless the reform h s direct, positive influence on these processes,
it
will
not improve learning and cannot be considered an educational success.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 173/375
256
eterminants of Student Learning
Lea rning is a dy na m ic proc ess that occurs over t im e; it is desc r ibed as a
ch an ge in an ind ivid ua l s kn ow led ge, skil ls or value s. The le arnin g of
interes t here is tha t which occurs w he n s tu den ts are expo sed to ins t ruct ion
and the curr iculum. I t is the type of learning that produces academic
achievem ent and tha t influences a s tu de nt s educat ion al a t ta in m ent an d
future occupa t ion. The m ore genera l norm s and values tha t s tu den ts learn
in school, while important for their personal development and social
respon sibil i ty, are no t of concern here .
Sorensen and Hal l inan 1977) p r esen t a m od el of the learning proc ess that
identif ies three key elements—abili ty, effor t and opportunit ies for learn-
ing— and that specifies the me chan isms tha t gov ern the re la t ionship am on g
these e lemen ts . The var ia tion am ong s tude nts in the am ou nt they learn in
school is determined by var iat ion in one or more of these three var iables .
The first element in the learning model is ability. Typically, ability has
bee n view ed as a stable, innate character is t ic tha t def ines a s tu de nt s
po tentia l to learn. Abil i ty ha s been conceived narro w ly, referr ing to a
specific set of cogn it ive proce sses, such as de du ctio n, inference an d m em o-
rization tha t are m eas ure d b y IQ tests . Kerckhoff (1993) po ints ou t that the
term abil i ty suggests continuity of performance and cer tain boundaries or
l imita t ions tha t a person c annot t ranscend.
Recen tly, theoris ts (e .g. G ardn er , 1993; Sternb erg, 1990) ha ve ex pa nd ed
the definition of ability to include such traits as creativity, logic, artistic
talent, and physical m anip ulatio n. At the sam e
t ime,
educators hav e achieved
a bet ter unders tanding of learning s tyles , recogniz ing tha t some s tudents
are pr im arily visua l learners , other au dito ry learners an d s ti ll oth ers learn
best thr ou gh physical involv em ent. Finally, research has sh ow n that abil i ty
is no t a fixed ch aracteristic b u t rathe r is influenc ed b y on e s con text an d
op po rtu nit i es to learn. As a result of these de ve lop m en ts in learn ing theo ry,
abil ity is no w seen as a bro ad , dy nam ic trai t that has both inhe ri ted an d
acqui r ed components .
Th e sec on d elem ent in the learnin g mo de l is effort. Like ability, effort is
determined by both individual and contextual var iables . Personal charac-
teristics that affect effort include need for achievement, parental pressure,
ambit ion, energy, and motivat ion. Among the contextual var iables tha t
affect effort are peer pressure, role models, social reinforcements, and
prev ail in g no rm s and va lues. Th us, var iat ion in s tu de nt learn ing is affected
by two broad se ts of individual var iables—those determining abi l i ty and
those determining effor t .
Besides the individual traits of ability and effort, the learning process is
compr ised of oppor tuni t ies to leam. S tudents cannot acquire knowledge
unless they are exposed to it , regardless of their ability and effort. The
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 174/375
57
am ou nt of inform at ion pre sented to s tud ents is de te rm ined b y the ins t ruc
t ional process . Teachers t ransm it informat ion to s tud ents , w ho in terna l ize
some degree of what i s presented. The na ture and content of the ins t ruc
t ional process affects the amount that s tudents leam.
These three basic characteris t ics—abil i ty, effort and opportunit ies for
learning— are the pr im ary de term inan ts of learning. Each of these com po
nents of learning is a function of other factors that affect their magnitude.
For exa m ple , a s tu de nt s social c lass , pa re nt s educ ation , race or e thnici ty,
gen der , and family back gr ou nd , affect a s tu de nt s abili ty level an d effort,
wh i le school charac ter is t ics , such as per- pup i l expen di tu re , organ iza t ional
differentiat ion an d s ize, an d teacher characteris t ics , such as m obil i ty, t rain
ing and pe da go gy , inf luence a s tu de nt s oppo rtuni t ies to learn . Var ia t ion in
one or m ore of these var iables pro du ces var ia t ion in the am ou nt a s tud ent
learn s . Differences acro ss s tu de nts in abil i ty, effort a nd o pp or tu ni ty exp lain
var ia t ion in learning across s tudents and over t ime.
The three pr i m ary de te rm inan ts of learning, abi li ty , e ffor t and op po rtu
nity, should be the ul t imate focus of al l educational reform efforts . These
variab les , an d the broa de r set of variab les that influence th em , m us t be the
target of change in a reform that aims to improve s tudent learning.
Change in Ability Effort
a n d
Opportunities for Learning
Change in abil i ty, effort or opportunit ies to learn produces change in a
s tu de nt s kno w ledg e . An increase in one or m ore of these de term inan ts of
learning should increase the amount a s tudent learns . However , i t i s
unlike ly that these variables ha ve an ad dit i ve effect on learning . An
addit ive model implies that high levels of abil i ty or effort can compensate
for few op po rtun i t ies to learn . In ac tua l ity , learning can not occur w i tho ut
the opportunity to learn, regardless of abil i ty and effort . Hence, a proper
specification of the relat ionship am on g these three elem ents of the lea rning
process i s an in terac t ive model . Sorensen and Hal l inan (1977) propose a
di f ferent ia l equat ion model to express the in terac t ion among these var i
ables and their effect on s tudent learning. An understanding of how to
increase abil i ty, effort and opportunit ies to learn wil l provide the founda
t ion for successful reform programs.
Ability
Since ability is , in part, inherited, it may represent the aspect of the
learning p rocess tha t is leas t am enable to change . Nev er the less , s i tua t ion al
factors are known to have an impact on abil i ty. Intel lectual s t imulat ion,
expo sure to kno wled ge , engagem ent in in te l lec tua l pursu i t s , and deve lo p
m en t of va rio us cog nit ive skil ls , increase cogn it ive functionin g an d e xp an d
and deepen the ability to learn. An intellectually rich family life, a stimulat-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 175/375
58
ing school atmosphere, opportunities to learn in and outside of school,
travel, and hobb ies, are likely to have a positive effect on a pe rso n s ability
and increase future learning capacity.
Kerckhoff (1993) argues that a reconceptualization of the concept of
ability
is
needed to avoid the self-fulfilling prophecy of stu dents labeled as
high ability prod ucing h igh achievement w hile those viewed as low ability
producing low achievement. A broader definition of ability and an
acknowledgement of the various dimensions of intelligence could be a
major factor in motivating stu den ts
to
achieve in areas that attract them and
in which they have talent and skills. This involvement in the learning
process could increase their ability to learn.
Effort
Effort is seen primarily as a function of motivation which, in turn, is
determined by personal characteristics and situational variables. Most
person al characteristics are genetically determ ined and remain fairly stable
over a lifetime. An in divid ual s curiosity, anxiety, sociability, and need for
achievement, approval, and status, are shaped in early childhood and
typically change little over time. Each of these characteristics affects h ow
much a person values learning, and thus how much one is motivated to
learn. The value attached
to
learning
is
function of the rew ards
it
provides.
Certain personality characteristics make learning more intrinsically re
warding to some students than to others. For example, those who are
intellectually curious or w ho hav e a high need for achievem ent, are likely
to derive greater satisfaction from learning.
For students who are less intrinsically motivated, an external reward
system can provide attractive reinforcements for studying. The compo
nents of a broadly based reward structure include tangible and intangible
rew ard s. These would include grades, aw ards, prizes, status, esteem a nd
social recognition. The mo re meaningful these rew ards are to a studen t, the
more effort the stud ent will expend to attain them . How ever, the accessibil
ity of the rewards is a critical factor in inspiring motivation and effort.
W hen grades are assigned only for level of achievement, less able stu dents
who are unable to attain high levels of achievement become discouraged
and abandon
the
effort
to
compete. When grades are assigned for g row th in
achievement, all studen ts have a chance to succeed an d the rew ard system
serves to motivate everyone to improve their performance.
The value a student attaches to status, esteem or social recognition,
depe nd s on the basis of the status system. If
school has a strong academic
orientation , then school autho rities, pa ren ts and s tude nts alike are likely to
value academic achievement and assign status to those who succeed
academically. The desire to attain status, then, is likely to motivate stud ents
to study in hopes of attaining adult and peer respect for achievement.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 176/375
159
However , if a s tudent at tends a less academically or iented school, or is a
member of a peer group tha t does not va lue academic achievement , then
oth er ba ses for sta tus w ill be de fined . These m ay h av e little or n o effect on
student motivation to learn or , indeed, may have a negative effect on
acad em ic effort . C on seq ue ntly , to increase a s tu de nt s effor t to learn, a n
external rewa rd sy s tem is nee ded tha t is both cons is tent w i th a s tud en t s
values and that provides social rewards for academic effor t .
Oppo rtunities to Learn
The mo st controllable e lem ent in the learning mode l is opp or tun i t ies to
learn. O pp ort un it ie s ar ise f rom three sources: the orga nizatio nal differen
t ia t ion of the s tudent body, ins t ruct ional pedagogy and the cur r iculum.
Each of these factors can be m an ipu late d in w ay s that have a direct im pac t
on s tudent learning.
a Organ izational Differen tiation O ne of the mo st notice able fea tures of
A m erica n schools is the w ay stu de nts are orga nize d for instruction . Dree ben
an d Barr (1988) state tha t schoo ls per form a series of transfo rm atio ns of the
composi t ion of the i r s tudent popula t ion by forming grades , c lasses , and
instruc tional gr ou p s w ithin classroom s. The aim of these transform ation s is
to increase the eff iciency an d effectiveness of instructio n. The mo st co m m on
basis for assign ing s t ud en ts to instructio nal units s age and abil i ty. Stud en ts
are assigned to grad es pr im arily on the basis of their age . Within grad e , they
often are assign ed to classroom s and other instructional gro up s on the basis
of abili ty. W ithin class abil ity gro up in g ten ds to be practiced at the elem en
tary level and across class abil i ty grouping or tracking is common at the
m idd le and secon dary level . O ther bases for ass igning s tud ents to gro up s
include gender , language abil i ty, and career interests . Increasingly more
po pu lar are hete rog ene ou sly g rou pe d classes , formed in the belief tha t
cogni t ive divers i ty can prom ote learn ing for a ll s tud ent s , par t icular ly those
of lower ability.
Research ha s no t revealed a direct effect of typ e of grou pin g o n s tu de nt
learning. However , methods of grouping s tudents have been associa ted
with var iance in instruction and pedagogy, which, in turn, affect s tudent
achievement. Kerckhoff (1993) claims that formal social structures, such as
abil i ty groups, systematically change s tudent character is t ics and alter
educat ional outcomes . He documents the cumulat ive ef fec ts of school
organ iza t ion on ach ievement and occupa t iona l a t t a inmen t , dem ons t r a t ing
significant long-term consequences of these effects.
In ad dit ion to grou pi ng effects on s tu de nt learnin g, the com po sit ion of a
group inf luences wi thin-group socia l processes tha t have an impact on
stu de nt effor t. Pu pils assigned to the sam e gro up d ev elo p a set of n o rm s tha t
govern effor t and level of achievement. Groups also form a comparative
reference whi ch affect a s tu de nt s acade mic self-confidence an d asp iratio ns.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 177/375
160
Th e greater a s tu de nt s self-confidence an d th e hig her the s tu de nt s aca
dem ic and career aspira t ions, the greater the motiv ation an d effort to learn.
The organizational differentiat ion of a school determines the composi
t ion of a s tu de nt s instructional un it . Co m po sit iona l characteris t ics affect
op po rtun i t ies to learn thro ugh the i r influence on ins t ruc t iona l ped ago gy
and the content of the curriculum. Student characteris t ics a lso engender
social processes that affect s tudent motivation and learning. Hence, oppor
tunit ies to learn are channelled through organizational differentiat ion.
b Instructional Pedag ogy The quant i ty and qual i ty of ins t ruc t ion rep
resent the mos t impo rtant opp ortun i t ies for s tud ents to learn . The g rea ter
the qua nt i ty of ins t ruc t ion and the h igh er i ts qua l i ty , the m ore informat ion
s tud en ts a re exposed to and the grea ter the ir oppo r tun i ty to increase the i r
kn ow led ge . Barr an d Dre eben (1983) ha ve sh ow n tha t the greater the rate of
instruction in reading groups, the higher the scores of f irs t grade s tudents
on rea di ng tes ts . W iley (1976) de m on stra ted that the long er the schoo l year,
the higher the achievement tes t scores . Heyns (1978) showed that s tudents
who rece ive more ins t ruc t ion by a t tending summer school s igni f icant ly
raise their achievement levels the fol lowing academic year.
Q ual i ty of instruction va ries by teacher characteris t ics an d ped ago gica l
techniques . Some teachers , by na ture , a re more engaging, authent ic , and
chal lenging to s tudents than o thers . Teacher t ra in ing programs can im
pr ov e a teach er s abili ty to com m un icate to a certain ex tent .
More amenable to change are the pedagogica l techniques employed by
teache rs . Tw o basic wa ys of instructin g s tu de nts are teacher-led instru ction
and s tudent-cen tered ins t ruc t ion . Teacher- led ins t ruc t ion is charac ter ized
by the teacher s be ing the center of com m unica t ion , prov idin g informat ion ,
d i rec t ing d iscuss ion , and moni tor ing s tudent input and behavior . The
pr imary modes of communica t ion inc lude lec tures , ques t ion and answer
sessions, reci ta t ions , and discussion. Proponents of teacher-led instruction
claim that i t insures that s tudents receive information in an orderly way,
and al lows mistakes to be corrected immediately. Cri t ics argue that s tu
de nt s are less interested in learning an d that they par t ic ipate less direct ly in
the instructional experience.
A n im po rtan t feature of teacher-led instruc tion is the potential im pac t of
the teacher s com m unica t ion pa t te rn s on s tu de nts opp ortu ni t ies to learn .
The interact ion of teachers and s tudents may be influenced by teacher
reac t ions to s tudents who vary in personal charac ter is t ics . Cons iderable
research dem on strates h ow teacher style differs by the gender, race, ethnicity,
an d, to a lesser degr ee, social c lass of the s tu de nt s (Fenne m a an d Pe terson ,
1987; G rieb and Easley, 1984; G oo d, 1970). Te ach ers tend to sp en d m or e
time with boys, ini t ia t ing m ore contacts with the m , correct ing, pra is ing , an d
encouraging them, whi le a lso soc ia l iz ing and joking wi th boys more f re
qu ently than w ith girls . As a result , boy s are instructe d in m o re s u p p o r t i v e
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 178/375
161
environ m ent, both academically and emotionally, than girls, w ho seem to
suffer benign neglect. Similarly, some studies show that white students
receive more praise for academic performance than black students who
receive more directives regarding behav ior. These differential com mu nica
tion pa ttern s affect stu de nt s attitudes tow ard school, self-esteem, achieve
ment and educational aspirations. By favoring one demographic group
over another, a teacher s pedagog ical technique increases opp ortun ities for
some students to learn, relative to their peers.
Student-centered instruction occurs primarily in peer instructional groups.
Stodo lsky (1984) identifies five types of peer instruction: completely
cooperative, cooperative, helping obligatory, helping p ermitted and peer
tutoring. Minor differences exist among these groups, but they all stress
stud en t initiative in the learning process and focus on individual and grou p
responsibility for learning. Students are seen as the primary resources for
learning, and s tude nt interaction
is
the
basic
mo de of instruction. Research
(Slavin, 1987; 1990) indicates that s tudents w ho w ork in peer instructional
gro ups attain higher dev elopm ental skills than those work ing alone, likely
a result of the discussion, argumentation and multiple perspectives that
characterize these gro ups . Peer instructional groups also have been shown
to imp rove social relations among stu den ts, particular those with different
backg rou nds or social characteristics. Im proved sociability leads to more
positive attitudes tow ard school, which likely affects st ud en ts motivation
and effort to learn.
c
Curriculum The content of the curriculum determines w hat is taught,
rathe r than ho w it is taugh t. The choice of curriculum content ord inarily is
determ ined by district and state educational authorities. Curren tly, there is
a mov e tow ard establishing a national curriculum, in conjunction with the
mo vem ent to w ard national assessment. Even in the absence of a national
curriculum, however, considerable similarity is found across school dis
tricts and states in the courses that are offered to elementary, m iddle and
secondary school students. Indeed, Meyer, Kamens and Benavot (1992)
argue that there is a general wo rldwid e consensus about the main elem ents
of the curriculum. The central elements of the curriculum seem to be
literature , mathematics, social studies, citizenship an d m orality. Most of the
cross-cultural variance in curriculum content stems from political and
cultural differences. Curricular variation in the United States is often
accounted for by such factors as the inclusion of regional geography and
history in social studies courses.
Most of the variation in curricular content that students experience is
attribu ted to differential placem ent in courses w ithin the school.
he
nature
of the curriculum to which a student is exposed is determined, almost
completely, by the track to which the student is assigned. In high school,
students take A dvan ced, Ho nors, Regular or Basic courses. The top three
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 179/375
262
levels of courses are college preparatory . They contain liberal arts, math
ematics, and science courses, as well as some skill courses, such as key-
boa rding and comp uter science. Students tend to use the same text book,
regardless of level, but the more adv anced studen ts cover more material in
the text, or use supplementary materials, and receive a faster rate of
instruction, than those in the Regular classes. More independent study is
incorporated into the higher ability courses as well. Students in the high o r
average tracks aim for an academic diploma on graduation, which has
somewhat rigorous course and sequence requirements. Students in the
Basic courses, on the other hand, take fewer academic courses, aim only for
a general diploma, tend to use more simplified text books, and take a
prep ond eranc e of vocational and applied courses. Consequ ently, the cur
riculum to which stu den ts in Basic courses are exposed is m arke dly differ
en t from their peers in the academic track.
Two impo rtant questions arise concerning the curriculum, m aking it the
target of educational reform. The first question is what changes in the
curriculum are needed to reflect changes in American society. One major
reform effort
is
directed toward including a multicultural perspec tive in the
curriculum that would act as a corrective for a prior overemphasis on the
influence of Western Europe on world change and development. The
Rainbow Curriculum (Agard-Jones, 1994; Bird,
1993
recently im plemen ted
in N ew York City is an example of this effort. This program was controver
sial from the start, representing values tha t conflicted with those of man y
pa ren ts, educa tors and other citizens. The Board of Education w as forced to
withdraw the program after only two years of trial.
The second question about the curriculum
is
whethe r it sho uld differ for
stu dents who vary in certain characteristics. Some school authorities favor
an academic curriculum for higher ability students to prepare them for
college, and a vocational curriculum for lower ability students to prepare
them for the labor market. Others argue that all students should receive a
basic liberal arts education, supplemented by vocational or other special
ized training. Moreover, failing to take certain college prep arato ry course s
at a particular grade level limits
a
student s
access
to future courses and may
preclu de their adm ission to many colleges.
To
insure that stud ents oppor
tunities to learn are maximized, students need experienced counseling to
guid e their course selection. They are best served by decisions that p rov ide
the greatest ongoing accessibility to the curriculum .
ducational Reform and Student Learning
In
1983,
the National Com mission on Excellence in Education, ap poin ted
by President Reagan, issued the influential report, ation at Risk This
report documented serious inadequacies in the American public school
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 180/375
63
system. A larmed by
the
findings of
this
report,
the
public became m otivated
to improve the schools. Several reform plans were formulated in the
subsequ ent d ecade. Some of these reforms succeeded in increasing stud ent
achievement, while others failed either to gain support or to improve
student learning.
The success or failure of any ed ucational reform is related to the goal of
the reform and the m eans employed to attain that goal. If the specific goal
of a reform is student learning, and the means used to attain that goal
involve pos itive interven tion in the learning process, the reform likely will
be successful. The U niversity of Chicago School Mathem atics Project
(1988,
1989) is just one exam ple. This curriculum was designed at the U niversity
of Chicago by a group of researchers and educa tors, aimed at impro ving
stud en ts ma them atics ability and achievement. The new curriculum pro
vided a conceptual approach to mathem atical concepts, required that the
students develop higher-order thinking
skills,
emphasized applications as
well as theory, and provided illustrations and exam ples that were salient to
the age and interests of the students being taught. The reform increased
stu de nt s mathem atical fluency, strengthened their belief that they could do
mathematics, and increased their opportunities to learn by an improved
curriculum and more relevant instructional techniques. As a result, the
reform is prod ucin g dramatic imp rovem ents in stud en ts mathem atics test
scores.
Even if
a
reform
has
a political or social goal, rather than an academ ic one,
if it relies on im prov ing learning to attain its goal, the reform is likely to be
a success. A p op ula r reform in the 1980 s w as the creation of m agn et schools.
n
addition
to the
regula r curriculum , magnet schools offered specialization
in the arts or sciences or some other area. The primary goal of this reform
was to increase the racial and ethnic integration of the schools on a
voluntary basis, in order to avoid forced integration through mandatory
buss ing. This goal was no t directly related
to
studen t learning. Ho wev er, to
attain integration, the schools needed to attract an ethnically and racially
diverse stud ent bod y. Consequently, the curriculum in these schools w as
carefully designed to meet the educational needs of a broad spectrum of
stud ents . The result was that magn et schools quickly became po pular and
attracted large enrollments. The goal of integration was accomplished by
improving opportunities to learn in these schools. The outcomes were
greater student involvement, and, as indicated by test scores, improved
learning.
On the other han d, if an educational reform has a non-educational goal
and em ploys m eans that have little to do with the learning process or that
hinder it, the reform is likely to fail in terms of student learning. For
exam ple, efforts to establish community-based control of schools failed in
many school districts, most notably Chicago and Brooklyn, New York,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 181/375
164
becau se the a im of the reform w as to chang e pow er bases . In a t tem pt in g to
achieve that aim , l i tt le at ten tion w as pa id to al tering educa tiona l p roce sses
in the schools.
Several s ignificant educational reforms presently are being advocated
and considered for adoption in public schools across the country. These
reform s includ e pa ren tal choice of schools ; perfo rm ance- based instru ction ;
na t ional assessment and a na t ional curr iculum; loca l school autonomy;
shar ed dec is ion-m aking for pare nts and teachers ; res t ruc tur ing teacher
t ra in ing p rog ram s, teacher cert if ica tion a nd profess ional mob i l i ty ; re quire d
exit examina t ions for grad uat io n f rom high school ; length ening the school
day and school year , and es tabl ishing year-round school ing; de t racking;
s tudent-centered ins t ruc t ion and coopera t ive learning; and redis t r ibut ion
of s tatew ide fund ing for educatio n. W hile al l of these chan ges are p ur po rte d
to improve educational outcomes, i t is rare that advocates of the reform
m ak e di rec t l inkages be tw een the pr op os ed reform a nd the learning pro
cess . Consequently, the considerable energy that often is expended in
pro m ot in g these pro gra m s often is m isdi rec ted to fea tures of the p ro gr am s
that have little or no direct effect on student learning. The goals of each of
these reforms must be c lear ly a r t icula ted and the way the innovat ion or
interv entio n w il l affect the learn ing process m us t be m ad e explici t in o rde r
to es t im ate th e l ikel ihood of the reform s success .
For i l lus t ra t ion , two contemporary educat ional re forms wi l l be exam
ined. The firs t of these reforms involves curricular innovation and the
secon d is the pla n to detra ck schools . The extent to w hich these reforms are
l ikely to increase s tu de nt abil i ty, effort or op po rtun it ies to learn w il l be the
basis of the analysis.
a Curriculum Reform Public respon se to societal change s in the U nite d
States ha s m ad e the cur ricu lum the object of nu m er ou s plan s to reform the
scho ols. O ne of the social forces th at acco un ts for the cu rre nt inte res t in the
cu rric ulu m is social reaction to the cha ng ing pop ula t ion co m posit io n of the
Un i ted S ta tes . The dem ogr aph ic characteris tics of th is coun try ha ve cha ng ed
dra m atica l ly in recent de cad es, resu lt ing in a shift from a fair ly la rge, w hi te
majori ty of Western European ancestry, to a smaller white majori ty and
increasin gly larger non-C auca sian, e thnic an d racial m inori t ies . This g reate r
pop ula t io n divers i ty has evoked pressure to bro ade n and divers ify the
curr ic ulum to m eet the langu age , soc ia l and academ ic ne ed s of the s tu de nts .
Minor i ty g roups have dem an ded a mo re mu l t icu l tu ra l cu r r i cu lum tha t
ackn ow ledg es an d credi ts the contr ibut ions of mino r i t ies to the gro w th of
Western civi l izat ion.
In New York City, the Rainbow Curriculum (Agard-Jones, 1994; Bird,
1993) w as one response to th is press ure . Educators dev elope d a nd ad op ted
a curr iculum tha t em pha s ized m ul t icul tura l i sm, c iv il l ibert ies , sex educa
t ion , an d lang uag e dev elopm ent . Since ser ious va lue conf licts rega rdin g the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 182/375
65
curriculum were never really resolved, the program failed to receive
community support and was terminated two years after it was adopted.
Consequently, its effect on student learning is difficult to eva luate. N ever
theless, the initial reaction of many minority families was positive. They
claimed that the new curriculum had greater relevance to their children
than the one it replaced and judged that it increased student interest in
school and their motivation to learn. In this case, a reform that was
politically motivated, promised to have a positive impact on learning
through its direct effect on student motivation and effort.
An other type of curriculum reform w as a response to the economic
recession of the
1980's.
Increased unem ploym ent, lower salaries and fewer
job oppo rtunities, coupled with
a
high drop-o ut rate from secondary school
and an increase in crime and violence among youth, focused attention on
schools as a means of solving these social problems. Until recently, the
schools had provided little direct training for the work world Kerckhoff,
this volume). Now , teachers, paren ts, stude nts, and emp loyers are begin
ning to examine the link between schooling, em ployment and income as a
m eans to economic growth and societal improvem ent.
s a
result, a num ber
of reforms were directed at preparing students for technical or service
positions in the labor
market.
These reforms aim
to
improve the curriculum
primarily for lower ability studen ts. Businesses are forming alliances with
schools to prov ide leadership, funding and expertise in training stu den ts
for en try level jobs in their organ izations.
Nearly every state in the nation is adopting reforms directed at increas
ing the marketability of non-college bound students. One such prog ram ,
entitled Take Schools that Work , was developed by the Sou thern Re
gional Education Board and has been implemented in nineteen states. It
emphasizes both academic studies and vocational-technical training. In
stead of being assigned to the highly disparaged general education track,
stud en ts involved in this program prepare for careers, college, or both. In
ma them atics, science, English and social studies classes, teachers add ress
real-life situations and app ly abstract concepts
to
concrete situations. Varia
tions on this program include post-high school education in a comm unity
college or technical school, in preparation for a career of the student's
choice. The school provides a warranty on the student's transcript that
attests to the stud ent's competence.
A similar and equally popu lar program is the Tech Prep curriculum that
already exists in
a
large num ber of
states.
The most recent state to adop t this
program is Indiana, in which the curriculum traditionally offered in the
basic track will be replaced by the Tech Prep curriculum beginning in the
1994-95 school year. Tech Prep is a performance-based prog ram designed
to provide students with essential academic and technical foundations to
enable them to gain employment or pursue further education after high
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 183/375
266
school gradua tion. It involves bus iness, industry and labor in the pla nnin g
delivery and evaluation of the curriculum. The Tech Prep curriculum is
designed around learning activities that emphasize problem solving, criti-
cal thinking skills, teamwork and cooperative learning skills. It also pro-
vides core courses in mathematics, science, language arts, economics, and
computer literacy.
W hile Take Schools that W ork , Tech Prep, and similar curriculum
innovations are fairly new, examination of the mechanisms that link the
goals of the programs to the expected outcomes provide insight into the
likelihood of their success. The prim ary goal of these reforms is to p rep are
students for the labor market, although they do include liberal arts courses
in the belief that these represent a useful know ledge base for e mploy ment.
The new curriculum is carefully designed w ith stu de nts' abilities in m ind
and is directed toward meeting their learning needs. It integrates abstract
and applied knowledge and reinforces what is learned through student
performance. The curriculum appears to represent a greater learning op-
por tunity for stude nts than that presented
in
the traditional general stud ies
curriculum. If appropriate pedagogy accompanies the improved curricu-
lum, it is likely that these curricular reforms will increase student effort.
M oreover, the em phasis on performance may increase s tud en ts' ability by
expanding the ways in which they learn. In general, educational reforms
such s these, that redesign the curriculum to better meet the learning nee ds
of stud ents, have considerable potential to promo te stud ent learning.
b Detracking A
second reform that is gaining considerable popular
support is the plan to detrack schools. Tracking, or the assignment of
students to courses on the basis of their ability, is practiced in mo st m idd le
and secondary schools, at least for mathematics and English instruction.
Relying on a subs tantial body of empirical research (Gam oran, 1986,1994;
Sorensen and Ha llinan, 1986; Oakes and Lipton, 1992), critics argu e that
tracking disadvantages low ability stude nts by providing them with infe-
rior instruction both in quality and quantity. Gamoran, in this volume,
points out that tracking has allocative and legitimizing, as well as socializing,
effects on students, all of which act to disadvantage the low ability pupil.
Reformists are urging school authorities to detrack schools, that is, to
replace hom ogene ous gro uping w ith heterogeneous group ing. Their ratio-
nale is that heterogeneous g rouping improves the instruc tion offered to low
ability students and provides them with a more challenging academic
climate. Several major educa tion grou ps, including the National Gover-
nors ' Association, the Carnegie Council on Adolescent Deve lopm ent, and
The College Board, have endorsed detracking as a means to improve
student achievement.
A second goal of detracking is to eliminate the racial and social class
segregation that accompanies hom ogeneou s grou ping. Given the current
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 184/375
267
correlation be twee n
race,
social class and academ ic ability, tracking resu lts
in a disproportionate number of white students being assigned to high
academic tracks and black and lower class students being placed in the
lower tracks. ince detracking creates mo re integrated instructional grou ps,
it has been endorsed by the NAACP Legal Defense fund, the ACLU, the
Ch ildren s Defense Fund, and the U.S. De partm ent of Ed ucation s Civil
Rights D ivision.
The success of a detracking plan depends on how the plan is imple-
mented. In some schools, tracking is simply replaced by heterogeneous
grouping, under the assumption that detracking automatically improves
students opportunities to learn. In these cases, teachers often are over-
wh elme d by the teaching dem and s created by a heterogeneously grou ped
class.
This is not surprising, since tracking originally was established to
facilitate instruction by reducing the diversity of student abilities. Anec-
dotal evidence indicates that simply detracking students without other
modifications in program and pedagogy produces little improvement in
learning and, in many cases, further obstructs the academic progress of
lower ability stud ents .
In other schools, detrack ing is accom panied by change in the culture of
a school. Oakes an d Lipton (1992) argue tha t tracking is em bedd ed in a set
of norm s and beliefs about how students learn and wha t they should learn.
These beliefs are built on closely held societal value s. For detrack ing to be
successful, it must be accompanied by changes in teachers attitudes and
expectations abo ut stud ent learning, including how learning occurs, how
diverse are student abilities, the extent to which learning ability can be
increased, and what knowledge is critical for students.
Besides being em bed ded in a set of norm s and values, current tracking
systems have political underpinnings. The political dimension of tracking
is associated with status differences, expectations and consequences for
academic an d occupational outcomes. Consequently, tracking is related to
the distribu tion of school resources, opp ortun ities, and credentials that are
valued in society. An un equa l distribution of these valued goods is particu-
larly troublesome when it encompasses race and class differences. For
detracking to be successful, school faculties need to identify and address
these political dimensions of tracking and reach a consensus that will
motivate the detracking innovation.
Little research is yet available that systematically exam ines the effects of
detracking on stude nt learning. Anecdotal evidence and an un ders tand ing
of the learning process suggest that simply replacing tracking by detrack ing
has little impact on student ability or effort, and does little to increase
students opportunities to learn . Detracking may achieve a political or social
goal, bu t fail in terms of improving studen t learning. How ever, if detracking
is accompanied b y changes in the curriculum and instructional p edag ogy,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 185/375
68
the likelihood of success is greatly increased. With improved curricu lum
and m ethods of instruction, opportunities to learn may be provided through
an integrated, thematic curriculum, cooperative learning approaches, a
variety of assessment techniques, long-term projects, use of teacher aid es,
and various other strategies that make the learning process more authentic,
interesting and relevant.
onclusions
The aim of most educational reforms is to better prep are stude nts for
their political, economic, and social roles in society. Reform plans tend to
m irror society s under lying values and reflect changes in that value system
over time. The most controversial reforms generally are responses to
conflicting values and conceptions about the role of school in society.
While it is nearly impossible to determine whether a particular educa
tional reform has an impact on one of the general public goals of schooling,
the impact of the reform on the intervening process of stud en t learning is
easier to determine. If a school reform facilitates or promotes student
learning, it is likely, in the long-run, to have a positive impact on more global
functions of education.
Reforms influence student learning wh en they are linked directly to the
learning process. If a reform expand s a stud en t s ability or capacity to learn,
or motivates a student to expend more effort in studying, then it has a
positive impact on the primary determinants of learning. If a reform
improves pedagogy, the curriculum, or the organization of students for
instruction , then it increases stu dents opp ortun ities to learn. y having an
impact on one or more of these factors, reforms can achieve the most
important purpose of schooling, namely, to improve student learning.
W hether reforms attain other political, economic or social ends , in a direct
or indirect way, is m ore difficult to measure. How ever, the likelihood that
an edu cationa l reform will attain one or more of these social goals increases
to the extent that the reform improves student learning.
N um erou s educational reforms currently are being prom oted in Am eri
can schools. It is critical to analyze each reform to determ ine w hether and
in what way the reform affects the learning process. Advocates of a
particular reform often speak eloquently about its lofty goals and about
how well the innovation is attaining those goals. But unless a reform
directly increases learning for all students, it cannot be called an edu cational
success, regardless of what other ends it attains. Global goals of education
are difficult to attain and the progress of a reform toward meeting those
goals is difficult to ascertain. It is both easier and m ore meaningful to define
the success of a reform simply in terms of whether it results in an increase
in student learning.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 186/375
169
otes
The author
is
grateful for funding for this research from the U.S. D epartment
of
Education, Office of Educational Research and Improvement, Grant #008610960
and from the Na tional Science Foundation, Grant #R117E1011139.
References
Agard-Jones, Leslie. 1994. Going over the 'rain bo w' curricu lum for mu lticultura l
education.
Education Digest
59(7):12.
Barr, Rebecca and Robert D reeben.
1983.
How Scltools Work. Chicago, IL: University
of Chicago Press.
Bird, W arren. 1993. Rainbow curriculum loses mo me ntum .
Christianity Today
37(4):68.
Bowles, Samuel and Herbert
Gintis. 1976. Schooling in Capitalist
Am erica New York:
Basic Books.
Dreeben, Robert and Rebecca Barr.
1988.
Classroom composition and the design of
instruction. Sociology o f Education 61:129-142.
Fennema, Elizabeth and Penelope Peterson. 1987. In Berliner, D.C. and B.V.
Rosenshine,
eds. Talk to Teachers.
Effective teaching for girls and
boys:
the same
or different? pp.111-125. New York: Random H ouse.
Gardner, 1993.
Multiple
Intelligences :
the theory in
practice. New York: Basic Books.
Gamoran, Adam. 1986. Instructional and institutional effects of ability grouping.
Sociology
o f
Education
59:185-198.
Good
Thom as L.
1970.
Which pup ils do teachers call on?
Elementary Scho ol Journal
70:190-98.
Good lad, John
1.1994.
A Place Called
School.
New York: McGraw Hill.
Grieb,
H. and J. Easley. 1984. In Steincamp, M . and M. L. Maehr (Eds.)
Women in
Science;
Volume
2:
dvances in Motivation and Achievement
A primary school
impediment to mathematical equity; can studies in role-dependent socializa
tion, Greenwich,
CT:
JAI Press, Inc. pp.317-362.
Hallinan, Maureen
T.
1994. Tracking: from theory to prac tice.
Sociology of Educa-
tion
67(2):79-84.
.. Further thoughts on tracking.
Sociology o f Education
67(2):89-91.
Heyns, Barbara. 1978.
Summer Learning and the Effects of
Schooling. New York:
Academic Press.
Kerckhoff
Alan C. 1993.
Diverging Pathways: Social Structure and Career
Deflections.
Cam bridge: Cambridge University Press.
Meyer John, David Kamens and Aaron Benavot, with Yun-Kyug Cha and Suk-Ying
Wong. 1992.
School Knowledge for the M asses: World Models and National Primary
Curricular Categories in the Twentieth
Century. Falmer Publishing Com pany,
Na tional Comm ission on Excellence in Education. April
1983. A Nation at
Risk:
The
Imperative for Educational
Reform. W ashington, D C , U.S. Government Printing
Office.
Oakes, Jeannie. 1994. More than misapplied technology: norm ative and political
response to Hallinan on tracking.
Sociology of Education
67(2):84-89. And One
more thought. p.91.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 187/375
27
Oakes, Jeannie and M artin L ipton. February 1992. Detracking schools: early
lessons from the field.
Phi Delta Kappan
448-454.
Slavin, Robert. 1987. Ability grouping and stud ent achievement in elem entary
schools: a best-evidence synthesis.
Review of Educational Research
57:293-336.
. 1990. Ability grouping in secondary schools: a best-evidence syn thesis.
Review of Educational Research
60(3):471-499.
Sorensen, Aage B. and M aureen
T.
Hallinan.
1977.
A reconceptualization of school
effects.
Sociology
of
Education
50:273-289.
.
1986.
The effects of ability grouping on growth in academic achievem ent.
merican Educational Research Journal
23(4)519-542.
Spring, Joel. 1991.
merican
Education: An
Introduction to Social and Political
Aspects
New York: Longman Publishing G roup.
Sternberg, Robert
J.
1990.
Metaphors o f the
Mind
Conceptions o f the Nature of Intelli-
gence
New York: Cam bridge University
Press.
Stodolsky, Susan.
1984.
In Peterson,
Penelope, Louise Cherry Wilkinson and Maureen
T.
Hallinan (Eds.)
The Social
Context of Instruction Framework for studying instructional processes in peer
work groups. Orlando : Academic Press, pp. 107-124.
University of Chicago School M athematics Project (UCSMP). 1989. K-6, Evanston,
IL: Everyday Learning Corporation. 1988. 7-12, Glenview , IL: ScottForesman.
Wiley, David. E. 1976. In Schooling and chievement in merican Society Sewell,
William H., Robert Hauser and David Featherman (Eds.) Another
hour
another
day : quantity of schooling, a potent path for policy. New
York:
Academic
Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 188/375
PART THREE
ducation and
Labor Force Linkages
The division of labor that has developed over the past two decades
between those w ho study schools and those wh o study the labor force has
left the linkage betw een those two institutional settings less well-analyzed
than it needs to be . The four p ape rs in Part III provide some indication of
the potential value of further work on this issue, although they go about
analyzing the linkage in very different ways.
In the first pape r, William Bridges seeks to clarify the role of edu cational
certification in determ ining w ork ers ' earn ings in the United States. To do
this,
he focuses on the impo rtance for wo rkers of credentialism, the
tendency of firms to reward the symbolic value of educational credentials
rather than, or in addition to, the actual trained capacities the credentials
represent. Using a broad conceptualization of types of industries an d an
analysis of pay levels associated with levels of educational attainm ent and
creden tials obtained , he shows that industries vary in these respec ts. He
concludes that credentialism is more often found in indus tries in which
the required skills are not obtained throug h formal training. In particu lar,
industries dominated by male workers (especially in the financial sector)
are more likely to provide rewards for the symbolic value of credentials.
Bridges then speculates about the future prospects of such add ed rew ards
in light of on-going changes in the labor force.
The second p ape r, by Thomas DiPrete and Patricia McM anus, also deals
with the earnings returns to education, but it does so within a com parative
framework. Building on earlier theory and research that shows a tighter
education-occupation linkage in Germany, they show that education as
such (once occupation is controlled) has a stronger effect on earnings in the
U.S. than in Germany, although even in German y direct effects of edu cation
are apparen t. They also show that earning s are m ore directly affected by
vocational credentials in Germany, while continued tenure with a single
employer
is
more imp ortant in
the U.S. s a
result of these differences in the
education-occupation linkage (tight in Germany, loose in the U.S.), earn-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 189/375
72
ings trajectories are also m ore uniform in G erm any . The variat ion s in the
U.S.
da ta a re cons is tent wi th Bridges observa t ions in the pre vio us pap er .
T h e s e t w o p a p e r s n o t o n l y h e l p t o c l a r i f y t h e i m p o r t a n c e o f
education-occupation l inkages, they also cast l ight on the varied relat ion
ships of both of those kinds of at tainment with earnings.
Th e oth er tw o pa pe rs in Pa rt III shift the focus from a very bro ad an alys is
of overall labor force patterns to very restricted portions of the labor force,
bu t the role of educa tiona l creden tials in the divis ion of labor an d in the l ives
of w or ke rs rem ain s in the forefront. In the first of thes e, Rob ert A lth au se r
and Toby Appel consider the two-t iered s tructure of several a l l ied health
occupations, the t iers being separated in ways often associated with dual
labor markets — dif ferent educat ion and credent ia l prerequis i tes , wage
differentials , blocked mobil i ty between t iers , and (theoret ical ly) a separa
t ion of tasks . The ad va nta ge s of the up pe r t ier w ork ers are legit im ized b y
bo th the i r supe r ior c redent ia ls an d the pr esu m ed task spec ia l iza t ion. H ow
ever, there is actually a great deal of ove rlap betw een the tasks per form ed
by w or ke rs in the tw o t iers . This raises the que st ion of w he th er (or ho w ) th e
two-t ie red s t ruc ture can pers ist un de r these condi t ions . Al th auser and
Appel specula te tha t i t s pers is tence may depend on other s t ruc tura l e le
m ents such as hospi ta l adminis t ra t ion s and professional orga niza t ions .
In the las t pa pe r in Part III , Barbara H eyn s also discusse s the ed uc atio n
of a particular part of the labor force, but her focus is on the institutions in
wh ich these prospec t ive work ers obta in the i r prep ara t ion . She analy zes the
backgrounds , va lues , a t t i tudes and expecta t ions of bus iness school s tu
de nt s after the fall of co m m un ism in Po land. In a perio d of swe ep ing
pol i t ica l and economic changes , the re la t ionship be tween educat ion and
labor force outco m es is at best prob lem atic . Edu cationa l pr og ra m s that are
clearly l inked to those cha nge s at tract s tud en ts w ith atypical at t i tu de s an d
values , and the tension between the old and the new can be seen in their
vie w s of the future — the co un try s and their ow n. H ey ns focuses espe cial ly
on the tens ion be tw een the ro le of w om en in the com m unis t sy s tem an d in
the newly emerging f ree market economy.
These four papers make i t c lear that the school-work l inkage is an
im po rtan t bu t a h ighly var ied on e , dep en de nt on the charac ter is t ics of the
social an d cultu ral con texts in w hich it is situ ated . Bridg es sh o w s differ
ences by ind us try in the labor force retu rns to edu catio n credentials w ithin
the U.S . , an d D iPre te and M cM anu s analyze such di fferences be tw een
societies . A ltha use r and A pp el describe instances in the U.S. in whic h
edu catio nal cred entials ap pe ar to be used to legit imate different levels of
pay even w he n the tasks perform ed are h ighly s imi lar. A nd, H ey ns uses the
case of Poland to sho w h ow societal cha nge m ak es it diff icult for s tu de nts
(and in this case especial ly women s tudents) to es t imate the best use they
can make of the newly established credentials they obtain.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 190/375
Educational Credentials
an d the Labor M arket:
An Inter-Industry Comparison
W
ridges
University of Illinois t Chicago
An ongoing deba te ex i s t s abou t why educa t ion enhances ind iv idua l s
success in the labor ma rke t . W itho ut sacrif icing the inheren t com plexity of
the issues , Bidwell (1989) provides a succinct s tatement of the range of
poss ib le answers :
have suggested that in present-day bureaucratic societies educational attain
ment has several m eanings. t may deno te componen ts of trained capacity for
job performance, ancillary bodies of information about the work world, or
secondary status characteristics. Each of these den otations, in turn , may be
substantive or symbolic, and the symbolic aspect may be either formal
(certificates earned) or informal. (133)
In th is pa per ,
inves t iga te how one symbol ic man ifes ta t ion of edu cat ion ,
credent ia l l ing , var ies across indus t r ia l labor marke ts in the US econ om y.
The und er ly ing pr em ise of th is ana lys is is that educ ation c an play a different
role in different labor ma rke t contexts . For exa m ple in som e set t ings, those
wi th m ore educa t ion m ay be m ore h igh ly rew arde d because of the t ra ined
capacity they ha ve acqu ired. In other s i tuat ions, those wit h hig her levels of
formal schooling may do better because of their greater exposure to and
knowledge of h igh s ta tus cul ture . Credent ia l l ing , the face t of educat ion
s tud ied her e , i s def ined as the preoc cup at ion of em ploy ers w i th em ploy ees
having crossed one or more educat ional s ta tus thresholds .
Al though a comple te ca ta log of the poss ib le ways in which educat ion
contributes to socio-economic success is beyond the scope of this paper—
and u nne cessary in l ight of Bidwel l s thorou gh analys is— a few w or ds are
in ord er concernin g som e of the m ain pos i t ions wh ich have been adv anc ed.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 191/375
74
The human cap i ta l /marg ina l p roduc t iv i ty theory r emains pe rhaps the
dominant explanat ion ( see Becker 1958 in economics , Featherman and
H au ser 1978 in sociology). Put s imp ly, formal schooling teaches ind ivi du
a ls product ivi ty-enhancing knowledge and ski l ls which are demanded by
em plo ye rs . (In Bidw ell 's f ramew ork, this view co rresp on ds to the cell
def ined by the intersection of the sub stan ce and traine d cap acity
con cep tual dimen sions.) The cultu ral capital theory of edu catio n asso ciated
w ith Bo rdieu 1977, D ima ggio an d M oh r 1985, an d oth ers , also ho lds that
educat ion is impor tant because of what it teaches, but in this case what is
imparted by schooling is familiarity with elite culture and the ability to fit
in at an ap pro pria te level in the larger s tatus orde r .
Co un terpo sed to these sub s tant ive view s of school ing are theor ies
wh ich s t ress the imp or tance of educat ion as a symbo l . Al th oug h Bidwel l ' s
scheme leaves room for s ix possibil i t ies here ( trained capacity vs . work
world information vs. secondary s tatus character is t ics each symbolized
formally or informally), not all of these possibilities seem very plausible
empir ically. In our view, meaningful dis t inctions can be drawn between
two types of symbolic theories: (1) those closest to human capital theory,
e.g. , s ignall ing or screening theory, in w hich m ore sc hooling is seen as a n
indica tor of som e latent desirable, b ut un obs erva ble, trai t
1
(Spence 1974;
Stiglitz 1975, Sp ilerm an an d L un de 1991); an d (2) tho se wh ich stress the
imp or tan ce of edu cat ion not as a symb ol of unobse rved prod uct iv e capac
i ty, but as a symbol of conformity to the s tructure of the educational
mer t iocracy itself
Symbolic theories of education are of ten, but not always, l inked to the
idea tha t c redent ia l ism produces discont inui t ies in the re la t ionship be
tw een edu catio n and socio-economic outco me s. W hat m atte rs is formal
recogn it ion of com pletion of blocks or integral units of ed uca tion al at tain
m en t. On ce again, Bidw ell 's analysis is informative. As occ upa tiona l
allocation mo ve s to the center of s tatu s allocation an d becom es increasingly
formalized with respect to evidence of trained capacity, the al location
sys tem tak es on a threshold structure in which occup tions re stratified
ccord-
ing to minimum educational requirements. (1989: 123 emphasis added) The
development of an industr ial-educational system in which the crossing of
normat ively def ined threshold points becomes valued in i t s own r ight ,
reg ard les s of w ha t it signifies in term s of train ed capa city, is the essence of
w hat, for the pu rpo ses of this pap er, will be labelled creden tialism.
A ltho ug h the emerg enc e of a credential system can be traced to en tran ce
req uire m en ts for well-established occ upa tional specialt ies , recent dev elo p
ments in labor market research suggest that a focus on occupations
per se
m ay be too limitin g. First, it is nece ssary to recogn ize tha t jobs (defined as
a bu nd le of tasks w ithin an organ ization) are of ten m ore mean ingful units ,
par t icular ly within internal labor markets , than occupations (Baron and
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 192/375
75
Bielby 1986; A ltha use r an d Kalleberg 1971). M any in div idu als o ccup y
posit ions and form identi t ies around organization-specif ic roles which
ha ve only vag ue co unte rpar ts in external labor m arkets (Wil l iamson 1975;
Bridges an d Villemez 1991) . Second, the enforcem ent of ed uca tion al
requ irem ents a nd the use of educ at ional sym bols in hi r ing, pro m otio n, an d
com pens at ion decis ions are , in varying d egrees , the preroga t ive of em ploy
ers an d their org aniz ation s. For exa m ple, a detailed an d interesting acco unt
of one em plo yer s s t ruggles wi th the adopt ion of m ore s t r ingent edu ca
t ional s tan da rd s is con tained in DiP rete s accou nt of the post-W orld W ar II
Fed eral Civil Service (1989,137ff). In su m , w h en a tten tion is focussed on th e
sym bolic me an in gs of edu catio n, it seem s logical to ask abo ut the ch aracter
istics of those parties, i .e. employers, who elicit, receive, and act on the
symbols in ques t ion.
Co nside r ing tho se symbol ic theor ies tha t see educa t ion as an indica tor of
latent prod uc tiv ity , i .e ., screening an d signall ing theories , one can m ak e an
addi t ion al d is tinc t ion. On one s ide are those app roach es whic h prov ide a
clear neo-classical , marginalis t account of the meaning of educational
a t ta inm ent . For exam ple in Spenc e s v iew, grea ter am ou nts of school ing
indicate the possession of productivity-related habits and trai ts which are
inherent ly unobservable . However , individuals wi th the des i red t ra i ts
confront low er costs in acquir ing the requisi te sym bol or s ignal , i .e ., mo re
yea rs of schoo ling, and acqu ire m ore of i t. Th us, em plo ye rs wh o select on
the bas is of m ore years of com pleted e duc at ion are choos ing indiv idua ls
w ho are less bo the red by the diff iculty of susta ined , disciplined a ttentio n to
m ore or less com plex cog nit ive tasks. I arg ue that this logic easily conform s
to a con t inuo us and sm ooth re la t ion betwe en socio-economic outco me s a nd
educat ional a t ta inment . That
is,
w hat ev er latent trai ts edu catio n is index ing
m ay v ery wel l be dis t r ibuted in the pop ula t io n in a re la t ive ly con t inuo us
fashion so tha t emp loye rs could ra tional ly discriminate be tw een an a ppl i
cant with one year of college completed and one with two years .
2
O the r theories of edu catio n as an indicator of a latent prod uct ivity trai t
tak e a sligh tly different turn . Berg for ex am ple states:
Education is often presumed also to be a continuous variable, approxim ately
the same m arginal differences in the economic values are assum ed to exist
betw een, say, any tw o successive years of high school. And researchers
assume this despite the recurrent finding that diplomas and degrees com
mand a price in the labor market that goes well beyond the marginal incre
ment of learning that may be achieved between the third and fourth year of
high school or college
(1971,
26)
And in explaining the reasons for this at tention to completed blocks of
school ing he repor ts on in terviews w ith bus iness leaders in w hich:
The college degree was consistently taken as bad ge of the holder s stability
and w as apparently
highly prized characteristic of young recruits. Most of
the responden ts ma de it perfectly plain that the content of the college program
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 193/375
176
mattered a good deal less than the fact of successful completion of stud ies.
(1971, 75)
In shor t , the use of educat ion as a symbol of unobserved product ive
capaci ty is cons is tent wi th both cont inuous and discont inuous reward
functions for years of completed schooling.
H ow ev er , bo th of these theories are different f rom a vie w of edu catio n
w hic h sees it less as a sym bol of latent pro du ctiv ity, an d m ore as a ma rk er
of deg ree hol de rs ' conformity to meritocratic culture.
T h u s ,
M icha el Faia in
Selection by Certification is atten tive to the possib ility tha t crede ntia lling
resul ts f rom i r ra t ional or non-econo mic behav ior on the pa r t of em ploy
ers and is explicit ly conc erned w ith locating discontinu it ies or ab ru pt shif ts
in labo r m ark et outc om es associated w ith the poss ession of formal cert if i
cates.
Figure 9 1 locates these different version s of the sym bolic m ea ni ng o
education along a continuum that ranges f rom those theories al leging that
education s ignif ies unobserved productivity to those al leging that i t s igni
fies cultural conformity.
Empirical Stu dies of reden tialling
M uch , bu t no t a l l, of the empi r ica l inves t iga t ion of edu ca t io na l
credential ism has been colored by a polemical tone in which the central
research a im is to discredit the re igning tech nica l / ra t ion al or h um an
cap ital vie w of the functions of ed uc atio n. Th is ideo logic al strain is rea dily
ap pa ren t in Col l ins ' 1979 volu me , The redential Society wh ere the open in g
ch ap ter is t i t led, The M yth of Tech nocracy. In pre sen ting an alterna tive
to the technolog ical function theory of edu catio n, Coll ins em brac es a
diffuse notio n of creden tial ism w hich d raw s wate r f rom the sam e well as the
related cultu ral cap ital ism arg um en t. Ed ucatio n m atters insofar as i t can
be interp reted as an indicato r of cultura l socialization w he the r or no t i t
inde xes sup erio r technical skills .
At the same t ime, Coll ins presents a s l ightly more nuanced view of the
issues. In par t icula r , he develo ps a set of em pir ical predic tions a bo ut the use
of educational requirements based on the notion that they play a key par t
in a l lowing organiza t ions to mainta in control through normat ive mecha
nism s ra ther than throu gh ins t rum enta l or coerc ive m ean s . A s a conse
quence , he examines whether educat ional requirements are higher in
pub l ic t rus t organiza t ions than ma rket or iented ones an d finds a pos i t ive
resul t. H e a lso inves t iga tes the cond i t ions un de r which bus in ess a dm inis
trat ion credentials are requ ired for prosp ective ma na ge rs finding th at these
de gree s are par t icu lar ly sou gh t after in, . ..(a) an extremely m ark etin g-
conscious manufactur ing group [food, fabr icated metals , pr inting]; (b) a
pub lic-relat ions-con scious tran spo rtation an d uti l i ties gr ou p [air an d wa ter
tran spo rtatio n, gas an d electric services]; an d (c) the adm inistra tors of large
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 194/375
77
FIGURE 9.1 Theories of Educational Symbolism
(Unmeasured (Cultural
Productivity) Conformity)
Theory Theorist:
Signalling/ Screening/ Ritualism/
Spence Berg Faia
What Education
Low Cost of Stability/ Con formity to
Symbolizes: Learning Stick-to-it- Cultural
iveness Norms
Relation Between
Education Reward:
Continuous Discontinuous Discontinuous
scale service org aniz atio n [hosp itals an d schoo ls] . (1979, 35) Finally, he
sugges ts tha t educat ional credent ia ls increase in impor tance when ant i
discr imin at ion law s m ak e over t ascr iptive and s ta tus-based se lec tion m ore
difficult.
3
For prese nt pu rpo ses , wh at i s imp or tant in Col l ins ' wo rk is not h is
def init ion of credential ism, which leans more toward cultural capital than
the edu cat ion al r i tua l ism a t i ssue here , bu t h is sugg es t ions abo ut bo th the
fact and sources of variation in credentialism across different segments of
the labor m arket .
A different, bu t no less po lem ical, stance is take n by M ichael Faia (1981),
al though the target of his cr i t icism is human capital and rational choice
econ om ics rath er tha n functionalis t sociology. U nlike Coll ins , ho w ev er , he
is no t par t icular ly in teres ted in exam ining var ia t ion am on g emp loye rs or
typ es of em pl oy ers in the use of edu catio n as a creden tial . In gene ral , his
f indings show that controll ing for years of schooling, those who hold
college and high schoo l dip lom as fare better in occ upa tional prestige term s
than those w i tho ut them. Once occupat ion is control led, how eve r , he
uncovers only minimal direct effects of certif icates on income.
In con trast , Sp ilerm an a nd L un de (1991) offer a relat ively non -ideo log i
cal as se ss m en t of th e role of forma l certificates.
4
They att em pt to find effects
of earn ed degrees which wo uld be cons is tent wi th wh at they term a var iant
of the 's ign al the sis ' (p 695). W hat they ha ve in m in d is that deg rees are
ov erw eig he d as s ignals a t the outse t of a h i r ing t ransact ion — me aning
that employers , in using cer t if icates , at tempt to measure latent productiv
i ty , bu t do so er roneo us ly . Thu s , they ini tia lly rew ard p eop le wi th degree s
beyond the level which they should, and later , employers correct their
mis tak es as the w ork er ' s t rue produ ct ivi ty becom es evide nt . H ow eve r ,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 195/375
78
their data d o no t bear o ut this sup po sit ion as they do not f ind the e xpe cted
negative interaction effects of possession of a degree and seniority on
pro m oti on rates , w hich is their de pe nd en t variable. Instea d, cert if icates
seem to exert an influence wh ich is m or e consis tent with the skil l acquis i
t ion thes is, than wi th the credent ia l ism thes is. F ur therm ore , becau se the i r
da ta al l com e from s ingle firm (and hen ce, single economic sector) , the y do
no t add res s the que st ion of contextual varia t ion in crede ntial effects .
The perspect ive on credent ia l l ing adopted here incorpora tes more e le
ments of the arguments presented by Col l ins and Faia than of those
prese nted b y Spi le rman an d Lu nde. L ike Faia, I am in teres ted in cap tur in g
a more pure ly symbol ic or ienta t ion toward educat ion than is impl ied by
Co l l ins ' proto-cul tura l -capi ta l a rgu m ent or Spi le rman an d Lu nd e 's ra t ional
choice arg um en t. W hile i t goes pe rh ap s too far to insis t that em plo ye rs w ho
require and reward completion of blocks of educational units do so for
irrat iona l reas on s, m y defini t ion of crede ntial ism d oes emp ha siz e that this
be ha vio r has a r i tual is t ic or cerem onial com po ne nt to i t w hich m ay n ot get
corrected as t ime passes . These aspects of org aniz ation al practice hav e bee n
repe atedly recognized in the recent pas t , beg innin g w i th the essay by M eyer
an d Row an (1977) an d culm inatin g in the so-called ne o inst i tu t ion al
perspective on organizational behav ior (DiM aggio and Powell
1983,
Edelman
1990, Zu ck er 1987).
However, l ike Coll ins , I am concerned with exploring variat ion in the
occurrence of credential l ing by examining different types of employment
set t ing s. The pu rp os e of this end eav or is f irst to tes t w he the r the use of
edu cat ion as a symbol ic c redent ia l is or is no t con stant across different pa rts
of the econ om ic land scap e. A ssu m ing th at it is not , the second goal is to
explore some pre l iminary hypotheses about what charac ter is t ics of eco
no m ic set t ings m ay contrib ute to this usa ge. Finally, i t is nece ssary to
inco rpo rate into the analysis , Bidw ell 's conception of blocks of e du catio nal
com ple t ion or educ at ional s ta tus thresholds . In descr ib ing ho w on e m igh t
ope rat ion alize the s ignall ing theory , he s tates , If the s ignall ing valu e of
ed uc atio n arises prim ari ly from the sym bolic aspec t of the thresh old s truc
ture , then the re la t ionship be tween educat ional and occupat ional a t ta in
m en t should a ppro xim ate a s ta ir -s tep function in wh ich the m ore pow erful
effects correspond to the transition points in the structure. . . (1989, 127)
A lth ou gh his readin g of s ignall ing theo ry is op en to deba te (see note 3), the
l inkage be tween educat ion as a symbol ic commodi ty and the s ta i r -s tep
functional form provides a useful tool for our subsequent analysis .
H y p o t h e s e s
The screening theory discussed above implies that a focus on blocks of
completed education wil l be more l ikely in skil led than unskil led jobs.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 196/375
79
H ow ev er, w ha t is cen tral to this theo ry is that the skil ls be those wh ich are
difficult to measure or determine at the ini t ia t ion of the employment
relat io nsh ip. Th us, ou r f irst hy po the sis is that ind ustr ies w ith a hig her
proport ion of jobs which require diff icult- to-measure skil ls wil l rely on
cert if icates rather than reward continuous years of schooling.
Imp lici t in som e cri t iques of the use of educa tion as a hir in g cri terion is
the idea tha t fixation on form al certif icates de va lue s the m ark et curre ncy of
those wi th mor e useful kno wle dg e obta ined in the rea l w or ld or on the job
its lf Follow ing this logic, m y next hy po the sis is that ind ustr ies with hig he r
rewards for long periods of work experience wil l be less l ikely to use
educat ion as a c redent ia l than those wi th lower rewards for work exper i
ence.
As suggested earl ier , the kind of credential ism at issue here is tanta
m ou nt to a r i tua l ist ic use of edu cat ion in em plo ym ent dec is ions . Fol lowing
a logic s imilar to tha t of Spilerm an a nd L un de
(1991), I
hypoth es ize tha t jobs
which require skil ls that are learned in school wil l tend
not
to reward
edu cat io n in a c redent ia l i st m ann er . Or s ta ted som ew hat d i f ferently , th e
m ore schoo l-based t ra ined capacity m at ters , the less f reedom em ploy ers
have to indulge their tas tes for the ceremonial , as opposed to the substan
t ive, aspe cts of e du catio n.
At tent ion to the ceremonia l m ean ings of edu cat ion is a lso qui te co m pat i
b le wi th the so-ca l led new ins t i tu t ional is t paradigm in organiza t ional
analy sis wh ich exp lains the ado ption of vario us em plo ym en t practices , e .g.,
pro m ot io n lad der s , equal opp ortu ni ty offices , person nel dep ar tm en ts , e tc. ,
as a t tempts to secure organiza t ional legi t imacy by adopt ing cul tura l ly
approved mode l s o f employment p rocedures . However , which o rgan iza
t ions are m ost likely to be influenced by w hich inst i tut ion al en vir on m en ts
de pe nd s la rge ly on the k ind of em plo ym ent prac tice un de r con s idera t ion.
There have been many s tudies of employment outcomes and prac t ices
ranging from due process guarantees (Edelman 1990), to salary class if ica
t ion s truc ture s (Dob bin, e t a l. 1993), to ge nd er an d racial inco m e disp ari t ies
(Beggs 1993) which have used 'proximity to the public sector ' as one
indicator of susceptibi l i ty to normative pressures . That is , legal and s tate
l inked press ures a re de term inat ive w he n the prac t ices in ques t ion sym bol
ize 'fa i rness '. H ow eve r , o ther governa nce s t ruc ture e lemen ts , in par t icular
in terna l labor m ark ets and personn el de pa r tm ent s , w ere found by Baron e t
al.
(1988) to be m or e prevale nt in service ind ustrie s , . . .wh ich de pe nd
heavi ly on proc ess ( ra ther than outcome) m easu res of perform ance , and ,
... formal ize op era t ing p roce dure s to increase the i r ow n perce ived legi tima
cy. The inst i tu t iona l pre ssu res in this case seem to arise less from ove rt
government in tervent ion, and more f rom conformity wi th soc ie ty-wide
no rm s abo ut d is t r ibut ive jus t ice . That is , if rew ard dis t r ibut ions c anno t be
just ified o n the basis of m eas ur ed differences in perfo rm anc e, the al locative
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 197/375
18
s t ructures which produce them can a t leas t be presented as procedura l ly
formalized and consis tent .
In the case of edu cation al c redential ism , will investigate sev eral sou rces
of ins t i tu t ional pressure . Th us , m y next hyp othes is i s tha t indus tr ies mo re
proximate to the public sector will be more l ikely to rely on educational
crede ntials than those fur ther rem ov ed. This predictio n is m otiv ated b y the
observat ion tha t edu cat ion itself at least in the U S, is closely tied to the state,
which serves as both a source of funding and a source of licensing for its
prac ti t ione rs . A n alternative insti tutiona l hy po thes is is that edu catio nal
crede ntials are m or e heav ily util ized by service ind ustr ie s and o ther sectors
with outputs which are di f f icul t to measure than by indus tr ies where
pro du cts are tangible or mo re eas ily me asure d, as in the manu factur in g or
utility ind us trie s. A final institutio nal hyp oth esi s reflects the finding s of
Col l ins per ta inin g to the de m an d for bus iness degrees in m arke t con-
sc iou s and publ ic re la t ions conscio us indus tr ies . That i s , I predic t tha t
industr ies which are more sensit ive to their public images will be more
attentive to having properly credentialed work forces than those less
a t tu ne d to these concerns .
Of course, bureaucracy itself has long been linked with the idea of
pro ced ura l formalism or red tap e and educa t ional credent ia ls a re easi ly
seen as a kind of unn ecessa ry formality. Berg, for exa m ple, spec ulates that
t he , no nra tion al use of formal cred entia ls . . . mig ht be taken as a s ignif icant
sym pto m of bureau pa tho lo gy . (1971 ,175)
s
Th us , m y next hy po thes is i s
tha t indus tr ies wi th grea ter concentra t ions of bureaucra t ic organiza t ions
will rely more readily on educational credentials .
Finally, there is the ma tter, w hich C ollins has also raised , of the d eg ree of
inc lus ion of minor i ty an d other grou ps his tor ica l ly dev alu ed in A me r ican
cu lture . T he pred iction here is that greater ope nn ess of an indu stry 's labor
force to black emp loye es and to female em plo yee s will be associated w ith
greater re l iance on educat ion al credent ia ls . Al th oug h this pro pos i t ion
could be defended in severa l w ays , the mo st com pel l ing arg um en t is tha t
de gre es, cert if icates , or t i tles pro vid e a basis of s tatus enha nc em en t w hic h
offsets the perceiv ed statu s deficiencies associated w ith bein g black o r female.
ata and Methods
ducational credentialling
The under lying premise of the argument presented above is tha t the
operation of a credential l ing regime will produce sharp discontinuit ies in
the function w hich relates rew ard s to edu cation , an d that these discon tinuities
wil l cor respond to cul tura l ly recognized s ta tus thresholds in the educa-
t ional system. There are at least two w ay s of looking for such discon tinuit ies .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 198/375
8
O ne is de r ive d from B idwe ll 's dep iction of the resu lt ing rew ard s truc ture as
following a s tai r -s tep function. The second is to search for rew ard shifts
associated wi th th e posse ssion of deg rees an d cer tif icates. A lth ou gh th e
f i r s t method motivates the procedures which are developed in th is paper ,
i t is worth consider ing the al ternative method in some detail s ince i t has
much intui t ive ly to recommend i t .
Cons ider f i r s t the arguments ra ised by Faia , who employs dummy
var iables for Re spo nd ent ' s Hig hes t Earned D egree in var iou s regress ion
analyses of occu pat ion and income on educ at ion. A rgu ing agains t us in g
thresholds in years of completed schooling he writes:
The correlations between primary/secondary schooling and high school
grad uatio n (.49) and between college schooling and bache lor's- or grad ua te-
level certification (.55 and .66, respectively) indicate that schooling and
certification variables are conceptually and statistically distinct, and that there
is substan tial da nger in using four years of college as a proxy for g raduation
(as in Jencks et al. 1979). Respondents holding the bachelor's degree as the
highe st degree have had anywhere from 14 to 20+ years of schooling. In
add ition , academic certification in Western
so ieties is
typically celebrated by
elaborate ceremonies amo unting to a rite of passage that rivals baptisms, bar
mitzvahs, initiations, engagemen ts, wedd ings, retirements, and funerals in its
social poignancy. (1981,1101)
Un for tunate ly , h is argu m en ts are no t as persu as ive as they f ir st app ear .
For ex am ple , w hile the correlat ions he repor ts seem low , the re are s tat is t ical ,
as well as substantive reasons for them not to be much higher .
6
More
impor tant ly , there are problems with his methodology which make the
effects of the hig hest de gr ee d u m m y variables diff icult to interp ret .
Essentia lly , h is pro ced ure is to in t roduc e the du m m y var iables for ed uca
t ional credentials into a m od el wh ich a lready includ es controls for years of
ed uc atio n (actu ally a no n-lin ear effect for yea rs of ed uc atio n). The effects
of the de gre e levels are then to pro du ce shif ts in intercep ts de pe nd in g on th e
hig hes t deg ree ea rne d. (See Figure 9.2) That this pro ce du re is op en to
varying interpre ta t ions can be seen more eas i ly by means of a s imple
com par ison as in Table 9.1. The differences in de gr ee sta tus after co ntro llin g
for years of schooling can be discerned by reading across the rows of this
table. O ne of the mea nin gs of ha vin g a high er or lower creden tial than
would be expected by the number of years of schooling is that one is
perceived as an over- or under- achiever for a given number of years of
school. Th us, rath er tha n ref lecting a fet ishism of deg rees on the par t of the
population of employers , a posit ive effect of degree s tatus controll ing for
ye ars of schoolin g m ig ht just as easily ind icate a preferen ce in favor of ove r-
achievers , e .g. , those who have achieved cer t if icates by spending fewer
years in schools compared to those who have fa i led to amass enough
grad uat io n credi ts . In shor t , a l thoug h ini tia lly plaus ible , the d u m m y
var iable meth od olo gy of Faia is not wi th out prob lem s.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 199/375
82
FIGURE 9.2 Faia's Model of Educationa l Creden tialism
Years of chooling
TABLE 9.1 Typology of Degree Recipients by Years of Schooling
Years of
Schooling
1 t o l l
2 to 15
16 or m ore
High School Degree
No
Dropouts
Not enough credits
Null?
Yes
Skipped grade/GED
Regular grad
Not enough credits
ollege Degree
Yes
Child prodigy
Accelerated BA
Regular grad
As an alternative to this approach, the method used here explici t ly
inco rpo rates the idea of a disc on tinu ou s, s tair -s tep function relat in g ed uc a
t ion to outcom e. To assess wheth er an indu s try is respo nd ing to creden t ia ls
rath er tha n con tinu ou sly vary ing levels of achiev em ent, the s tair -s tep
functional form is directly compared to other functional forms and i ts
superior i ty, or lack thereof is evaluated by means of a test of statistical
s ignificance. In other wo rds , the questio n is not wh eth er the th resh old
cre den tial m od el fits; it is ho w w ell it f its co m pa red to the likely alte rna tive s.
In Figu re
9.3
and Figure
9.4 1
show the credent ia l m od el
I
ha ve specified and
two of the alternative models it is tested against.
The alternative m od el in Figure 9.3 is a mo dif ication of the bil inear spline
function uti l ized by Featherman and Hauser to capture differ ing returns to
edu catio n at different levels of the ed uca tion system . Their interpreta tion
of a s imilar mo de l, one tha t includes different s lopes for educ ation before
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 200/375
83
FIGURE 9.3 Stair-step Creden tial v s. Spline Functions of Educational Reward
7T
• o
to 5
| 4 t
Continuous splineMunction
Credential Stair-step function \ /
10
Years of Schooling
FIGURE 9.4 Stair-step Creden tial Models with Correct and Low Thresholds
10 15
Years of Schooling
20
an d after gra de 1 2, is that . .. the ad dit io nal pr em iu m for college is consis tent
with s tan da rd econo mic in terpre ta t ions tha t the col lege-educated are mo re
pro du ctiv e or trainab le tha n others w ork ers . (1978, 261) Th us, this
cont inuous , p iece-wise regress ion model can be taken to represent the
h u m an capi ta l or technologica l funct ion theor ies w hich are an ath em a to
Collins and Faia.
A few moments ref lec t ion, however , should convince the reader tha t
these tw o m od els are not easi ly com pare d to one another us ing conv ent ion-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 201/375
84
al statistical tests. Th at is
each
of the al ternativ e mo dels con tains term s
which are not inc luded in the o ther—a s i tua t ion known as non-nes ted
altern ative m od els in the s tatis tical l i terature . Recently, ho w ev er, so m e
app rop r ia te m etho ds for th is s i tua t ion hav e been bro ugh t to the a t tent ion of
sociological aud ience s . (Weakliem 1992; H alab y and W eakliem 1993) Build
ing on earlier statistical w ork (Cox
1961,
Pe sara n 1974), they in trod uce tes ts
which can be used to de termine whether each of the two non-nes ted
alternatives s ignificantly outperforms the other. In fact , Weakliem (1993)
pres ents tw o dif fe rent tes ts which he de m ons tra tes to be asym ptot ica l ly
equiva lent , the "N tes t" , der ived by Pesaran and used in Weakl iem and
Halaby 's income models , and the "P tes t" , developed by Davidson and
M ack inn on (1983). W ith ei ther of these tests , four basic ou tcom es a re
possible: (1) the al ternative model—the spline regression function in this
instance— can reject, i .e ., outp erfor m the nu ll m ode l— the cred entia l mo de l ,
but not vice-versa; (2) the null model can reject the alternative, but not vice
ver sa; (3) bo th can m ut ua lly reject each other; and (4) ne ith er can reject the
other. In this analysis , I ru n separate incom e regressio ns (described below )
within each of the Census detai led industries and identify those cases in
w hic h, as in (2) abov e, the credential ing m od el rejects the spline m od el . Th e
test us ed he re is the P tes t wh ich is so m ew ha t less com pu tat ion ally intens ive
than the N tes t .
7
There a re two other a lte rna t ives aga ins t wh ich the c redent ia l l ing m od el
of Figure 9 .3 is eva luate d, one of w hich is sh ow n in Fig ure 9 .4. Th e
alternative mo de lh as the same basic , discon tinuou s form as the creden tial ling
m od el , w i th one im po rtan t d i f fe rence—the threshold p oints a re loca ted on e
grade level below the cultural ly defined thresholds of 8,12, and 16 years
use d in the creden tialling m od el. The logic of this test is that the cr ede ntiallin g
m o d el is m ore th an a s im ple test of disco ntinu ity in a rew ard function, b ut
incorpora tes the added res t r ic t ion tha t the d iscont inui ty mus t occur a t
cultu ral ly me aning ful p oin ts . Ag ain, cases in wh ich the nu ll m od el rejects
the al ternative, but not vice-versa, are taken as evidence in favor of the
creden t ia l hyp othes is w i th in an indus t ry . A th i rd se t of com parison s i s a lso
m ad e w hich are s imilar to those of Figure 9.4 except that the cutp oin ts are
loca ted one g r ad e l eve l abo ve the th r e sh o ld s incor po ra te d in the
credent ia l l ing model .
Before describing the data which are used to es t imate these models , I
should ment ion some of the weaknesses and l imi ta t ions of the technique
use d here. As Bidw ell as tutely po ints out , the thresh old s tru ctu re of
A m erican e duc ation is not as s imp le as w ha t is rep resen ted in these f igures .
The m os t im por tant compl ica t ion , wh ich has been impo ss ib le to in t ro duc e
here, is that the system of edu cation al credentials differentiates horiz onta l ly
as one moves progress ive ly h igher beyond secondary educat ion and col
lege. Th is is part icu larly true as one conside rs the interrelat ion of edu ca-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 202/375
85
t ional c redent ia ls and occupat ional entry require me nts . Ho we ver , the topic
here is be tw een - ind us try var ia t ion in educat ional ly credent ia l l ing and the
fact that occupation-related credentials are not fully detected here is not
par t icu lar ly trou blin g. A second pitfall is that there is som e sl ip pa ge
between years of completed educat ion and the possess ion of degrees and
di plo m as, al tho ug h no t near ly so m uc h as Faia claims. Because of l imita
t ions in the data us ed , som e error is inh eren t in est im ating the pres enc e of
credential effects.
8
Data Edu cation Regressions W ithin Industries
The basic source of data for the within- industry regressions is the 1%
Public Use Mic rod ata Sa mp le (PUMS) from the 1980 US C en sus of Pop ula
t ion. The ad va nt ag e of this sam ple is that it con tains en ou gh obs erva tions
to pro vid e a sui tably large nu m ber of cases wi thin each indu s try . These data
have been repeatedly used in other labor market s tudies and provide
measures of many of the var iables which are necessary for present pur
poses .
Because the subseq uen t analys is wi l l involve com par ing the resul ts
of significance tests across different industries, samples of a constant size,
i.e., 1500 ele m en ts, w er e dr aw n from ea ch of the 3-digit ind us trie s identified
in the Cen sus .
9
Ind ividu als we re exclude d f rom the ind us t ry regress ions if
(a) they w ere not w ag e and salary wo rke rs; (b) their ind us try aff il iat ion w as
not rep orte d; or (3) they w ork ed less tha n 1 w eek in 1979.
The fol lowing var iables were extrac ted and, where necessary, recom
pu ted for use in the income regressions . The dep en de nt var iable w as w age
and salary income in 1979. The individual 's total work hours in 1979 were
es t imated by mul t ip lying the usual hours worked per week t imes the
nu m be r of wee ks wo rked in the year . Three w ork exper ience variables we re
use d in the basic regression s. T he f irst two were der iv ed from the s ta nd ard
transfo rma tion us ed to est imate w ork expe rience from C ens us data , i.e . the
ind iv id ua l ' s age m inu s h ighes t g rade a t t ended m inus 6. Th is wa s en te red
in qu ad ratic form as experience and experience squ ared . The third w ork
exper ience m easu re was a w ork in te r rup t ion du m m y var iab le based on a
ques t ion about the respondent ' s work ac t iv i ty f ive years ago (a ques t ion
use d in the 1980 C en su s, b ut not repe ated in 1990). Individuals with at least
five poten tial ye ars of w ork experience w ere scored 1 on this d u m m y
variable if they w ere no t wo rkin g in 1975, all oth ers we re scored z ero. Sex
wa s c ode d if resp on den ts were female and 0 if they were male . D um m y
var iables we re a lso inc lud ed for race ( l=black) and Sp anish or igin ( l=M exi-
can, Pu erto Rican, Cu ba n, or O ther Spa nish) . In add it ion , control var iables
were added for abil i ty to speak English (0=only speaks English to 4= Not
at al l ) , an d for no n-m etro po li tan residence ( l= ou tsid e SMSA).
Education was, of course, treated in varying ways in the different
specif ica t ions descr ibed above. Al though each equat ion inc luded a te rm
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 203/375
86
rang ing f rom 0 to 20 corresp ond ing to the resp on den t ' s h ighes t gra de
comple ted , the meaning of th is te rm var ies depending upon the o ther
ed uc atio n term s inclu ded . In the spline, or var yin g s lopes, specificat ion, the
coefficient for the overall education term indexes the slope of income on
ed uc atio n in the lowe st edu cation gr ou p, 12 years or less. In the c red ential
or sta ir- ste p specifications, the coefficient for the overa ll ed uc ati on ter m
indexe s the change in income per year of educa t ion w i th in each thresho ld
block, w he re the s lopes are assu m ed cons tant .
Data: Credential Regressions Across Industries
To inves t iga te our hypotheses about the varying use of educat ion as a
credent ia l across em ploye rs , several var iables nee ded to be mea sure d a t the
indu strial level . Firs t, each ind us try 's labor force was char acterize d by tw o
demographic var iables : the propor t ion of female employees and the pro
po rt io n of black em ploy ees. These w ere calculated from the sam e sa m ple
used for the wi th in- indu s t ry regress ions .
Two var iables represent ing rewards for work exper ience were der ived
from the within-industry regression coefficients of income on experience
an d exper ience squa red. A l thou gh the genera l form of the income-ex per i
ence functions were the same in al l industries , the familiar inverted U-
sh ap e, the specific form s of these functions differed in tw o wa ys acro ss the
ind ustr ies un de r con siderat ion . First , som e we re flatter than othe rs , indicat
ing a low er total retu rn to experience over the life course . (A m ain ten et
of h u m a n cap ital theory is that female w ork ers are m ore l ikely to be fou nd
in w ork se t t ings which offer less retu rn to experience.) The cu rve s also
differed in a second wa y, ho w ev er, w hich wa s the location of the tu rn in g
po int in the incom e retu rns to experience. Th at
i s,
in som e ind ustrie s skill
de pr eci at io n sets in at la ter ages than in othe r ind ustrie s . To the exten t that
edu catio nal credentials favor the yo un g over the old, we wou ld ex pect th at
t h e t u rn i n g p o i n t m e a s u re w o u l d b e n e g a t i v e l y c o r r e l a t e d w i t h
credential ism across industries .
1 0
It is important to keep in mind that the
dependent variable at issue in this paper is not s imply the return to
educat ion, but how much be t te r a d iscont inuous educat ion spec i f ica t ion
per fo rms com pared to a con t inuou s one .
Requ ired skil ls for each indu s t ry w ere aggregated f rom two CPS sup ple
m ent in terview s which w ere conducted in Janua ry of
1981
an d Jan uar y of
1983.
Each em ploye d re spo nd ent wa s asked to indica te wh ethe r the job
which he or she now held required skil ls obtained outs ide of the current
em plo ying o rganiza t ion. The two var iables used here a re 1) the pro por t ion
of respo nd en ts wh o said that their jobs req uire d skills obta ined from formal
com pan y t ra in ing in a prev ious pos i t ion , and (2) the pro por t ion of resp on
de nts w ho said that their jobs required skills obtained from forma l scho ol
ing . If cre de ntia ls are bein g us ed as an aid in the pro cess of selecting
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 204/375
187
employees into jobs where skills are required, these variables would be
expected to be associated with credentialling in a positive direction. On the
other hand , if credentialling is tantam oun t to a ritualistic use of educa tion,
these variables would be expected to be negatively associated with it.
11
To investigate the three institutional hypotheses, several other variables
were taken into account. For some institutional variables, da ta were
extracted from the US Com merce D epartm ent, Bureau of Economic Analy
sis file, The Detailed Input-Ou tput Structure of
the US
Economy, 1977 . To
cap ture the influence of consum er market orientation with in each indus try,
advertising expen ditures in millions of dollars by the indu stry were includ
ed.
12
For the second dim ension of labor market institutionalization, suscep ti
bility to gov ernm ent influence, the percentage of all purchases which were
m ade by governm ent bodies from each industry were used as an indicator.
13
In add ition, a du m m y variable coded
I
for industries in the Public
Adm inistration sector was included. As a first approximation to repre
senting the institu tional factor of having a non-tangible product, dum my
variables for the Finance , and Service sectors were also included as
predictors. Finally, as a rough indicator of industry bureaucratization, a
m easure of average firm size in each industry was utilized.
14
Findings
Table 9.2 sum m arizes the results of the three separate non-nested m odel
tests across the 23 industries included in this
analysis.
As noted, four types
of outcom es are possible w ith each of these tests. Evidence in favor of the
credentialling m odel occurs when the comparison m odel
is
taken
as
the null
hyp othesis, and the credentialling model rejects or outperform s it Indus
tries with this outcom e are classified in category (row) 4 of this table. The
opposite case exists when the alternative model performs better than the
credentialling model, and these outcomes are tabulated in row 1 of Table
9.2. Inconclusive results occur when both or neither m odel rejects the other.
Industries in one of these situations are tabulated in row s 2 and 3.
As expected, the credentialling m ode l does not apply to all segm ents of
the labor market. Depending upon the alternative model considered , it is
significantly better in between
26%
and 53% of the indu stries. As a test of
the agreem ent am ong the different tests of credentialling, c ross-tabulations
were constructed on these three variables after the two middle outcomes,
i.e., those repre sen ting inconclusive results, were combined, (see Table 9.3)
Although the re is no t perfect correspondence am ong these different tests,
the level of agreement is reasonably h igh.
Before attempting to explain this between-industry variation in educa
tional creden tialling, it is worthwhile exam ining the actual income reg res
sions from two industries that differ on these dim ensions. Sample within-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 205/375
188
TABLE 9.2: Outcomes of Non-nested Tests of Threshold-Credential M odel
Across 230 Industries: US Census D ata 1980
1
2
3
4
Outcomes of
Non Nested Tests
MODELS BEING COMPARED
Credential vs
low threshold
model
of
Industries
Alternative model
rejects credential model 7.4
Neithe r model rejects
other
Both models reject
other
Credential model
rejects alternative
Total*
Number of cases
36.1
3.9
52.6
100.0%
(230)
Credential vs
high threshold
model
of
Industries
17.0
13.5
35.2
34.3
100.0%
(230)
Credential vs
differing slope
model
of
Industries
37.4
20.0
17.0
25.7
100.1%
(230)
Totals do not add to
100%
because of round ing
ind us try regressio ns are sh ow n in Tables 9.4 an d 9.5. Results from a
credent ia l ized in du s t ry , Pu lp , Pap er and Pa perb oard M ills , a re sh ow n in
Table
9.4.
(F ollowing earl ier work , e .g. Ha laby an d W eakliem , a cutoff valu e
of 2.50, ra th er than 1.96,
is
us ed as a criterion for rejecting on e m o de l in favor
of anoth er. ) This ind us try satisfies al l of the three no n-n este d tests sh ow n in
Table 9.2: The P s tatis t ic for the creden tial m od el co m pa red to the
differing slope model is 5.85 (the reverse test has a value of 1.76 ; for the
credent ia l m od el com pare d to the low threshold m od el the s tat i st ic is 5.60
(the reve rse test ha s a valu e of 0.04); an d for the creden tial mo de l co m pa re d
to the hig h thr esh old m od el, the valu e of the s tat is t ic is 7.70 (the reverse
test ha s a va lu e of 0.84). C on siste nt w ith exp ecta tion s, the effect of th e
overa l l years of edu cat ion var iable near ly d isapp ears w he n the credent ia l
disco ntinu it ies are taken into accou nt in m od el 2 . (The difference in s lope
and di fference in in tercept param eters hav e been cod ed in such a w ay tha t
the coefficients for each term represent the
change
in the size of th e effect
from the pr ev iou s level . In M odel 2 , for exa m ple, there is an up w ar d shift
of $1700 in ear nin gs at 12 yea rs of schooling an d a no ther u p w a rd shift of
$8500 at 16 ye ar s of scho oling .) Th e effects of the oth er v aria ble s in the
m od el a re a lso la rgely cons is tent w i th expecta t ions . There are wag e
penal t ies associa ted wi th be ing female and wi th be ing Black—although
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 206/375
89
TABLE 9.3: Co nsistenc y of N on-n ested Tests of Th resh old-C rede ntial M od el
Ac ross 230 Indu stries: US Ce nsu s D ata 1980
Alt . model
rejects Inconclusive Cred. m od el
cred . m od el resu lt rejects alt . Total
Low Threshold
as Alternative High
Threshold
as Alternative
Alt. model rejects
crede nt ial m od el 41.2 35.3 23.5 100.0
(17)
Inconc lusive resu l t 26.1 47.8 26.1 100.0
(92)
Credential model rejects
alte rna tive 6.6 51.2 42.2 100.0
(121)
Gamma =.412, p<.001
Low Threshold
as Alternative Different
lope
Model
as
Alternative
Alt. model rejects
cre den tial m od el 100.0 0.0 0.0 100.0
(17)
Inco nclu sive res ult 45.7 45.7 8.7 100.1
(92)
Credential model rejects
alte rna tive 22.3 35.5 42.2 100.0
(121)
Gamma =.669, p<.001
High T hreslwld
as Alternative Different lope Model as Alternative
Alt. model rejects
credent ial m ode l 82.1 15.4 2.6 100.1
(8)
Inco nclu sive res ult 34.8 52.7 12.5 100.0
108)
Credential model rejects
alter nativ e 19.0 25.3 55.7 100.0
(114)
Gamma=.686 p<.001
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 207/375
190
TABLE
9.4:
Within Industry Regression in Sample Creden tial Industry:
Pulp,
Paper, and Paperboard Mills Census Industry 160 (Num ber of Cases=1500)
Model Model 2 Model 3
b t b t b t
Intercept
H o u r s
Worked
Experience
Experience
Sqrd
Work Inter
rup t ions
Black
Latino
English
Problems
Female
N o n - m e t r o
Educat ion
-15786.0
6.46
585.1
-8.22
-2445.1
-2996.3
-257.7
-1048.5
-5189.2
-435.4
1311.4
In tercep t Changes
At 12 Years
At 16 Years
-10.42*
16.74*
11.38*
-7.20*
-2.69*
-3.91*
-0.13
-1.74
-8.69*
-1.00
13.46*
-5664.3
6.00
615.0
-9.52
-2746.1
-3484.5
-905.2
-1700.6
-4966.2
-113.4
362.5
1745.3
8525.6
-2.80*
15.72*
12.21*
-8.44*
-3.08*
-4.64*
-0.46
-2.85*
-8.49*
-0.26
2.00*
2.19*
8.28*
-7938.8
6.23
632.8
-9.84
-2736.4
-3350.9
-1047.0
-1612.2
-5034.4
-144.2
601.2
-4.08*
16.28*
12.33*
-8.50*
-3.04*
-4.42*
-0.53
-2.67*
-8.53*
-0.33
4.02*
Slope C hanges
At 2 Years 1306.5 5.12*
At 6 Years 905.2 1.19
R sqd Ad) .441 4® .455
*t significant, p<.05
none with being Lat ino—and earnings are lower for those wi th work
inter rupt ions , those wi th di f f icul ty speaking Engl ish , and those working
fewer hours .
Table 9 .5 shows the wage pat terns for the Membership Organiza t ion
ind us try in wh ich the credent ia l m od el does no t uniq uely ou tper fo rm the
altern ative m od els on any of the no n-ne sted tests . Th e P s tat is t ic for the
cred ential m od el com pa red to the differing s lope mod el is - .60 ( the reverse
test ha s a valu e of 6.71); for the credential m od el co m pa red to the low
thr esh old m o d el the statistic is 1.51 (the rev erse test ha s a va lu e of 2.11 and
neit her m od el rejects the other on this test); an d for the crede ntial m od el
co m pa red to the high thre sho ld m od el, the valu e of the s tat is t ic is .34 ( the
rev erse test h as a va lu e of 4.50). Th ere are also so m e sligh t differences in the
effects of many variables. The effect of poor English skills is not significant
in this ind us try, an d there is no s ignif icant de crem ent in earn ing s associated
w ith Latino ethnicity, with being Black, or with h av ing a w ork inte rrup tion .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 208/375
9
TABLE 9.5 Within Industry Regression in Sample Non-Credential Industry:
Mem bership Organizations- Census Industry 88 (Number of Cases=1500)
Model Model 2 Model 3
b t b t b t
Intercept
H o u r s
Worked
Experience
Experience
Sqrd
Work In ter
rup t ions
Black
Latino
English
Problems
Female
N o n - m e t r o
Educat ion
-6505.9
5.27
389.0
-6.33
-1185.3
-1098.7
-236.5
687,4
-5361.4
-2244.7
757.6
In tercep t Changes
At 12 Years
At 16 Years
-4.75*
16.97*
8.98*
-6.95*
-1.72
-1.43
-0.13
1.17
-11.09*
-3.50*
8.31*
-7334.3
5.31
398.9
-6.51
-1176.8
-1192.5
-282.4
662.2
-5180.0
-2222.6
900.1
-1524.9
-78.6
-3.67*
16.96*
9.15*
-7.11*
-1.70
-1.55
-0.16
1.13
-10.51*
-3.46*
4.88*
1.72
-0.08
-2290.0
5.22
386.7
-6.51
-1142.7
-1123.0
117.7
370.8
-5074.6
-2170.8
414.7
-1.00
17.03*
8.97
-7.13*
-1.68
-1.48
-0.078
0.64
-10.52*
-3.43*
2.12*
Slope C hanges
At 2 Years -148.6 -0.53
At 6 Years 2868.9 6.35*
R-sqd Adj .435 .436 .452
*t significant, p<.05
Living in a non -me tropo l i tan se t t ing does , how ever , d iminish o ne s income
in this industry.
While these regression results i l lustrate the differences between indus
tr ies with greater and lesser degrees of credential l ing, they also introduce
a cautionary note into the subsequent analysis . The non-nested tests here
detec t re la t ive ly sup er ior per formance in a m od el even w he n the u nd er ly
ing difference in va rian ce expla ine d is very m od est, (see Tab les 9.4 an d 9.5)
T h u s ,
a t this phase of the analysis , we conclude that educational creden-
tial ism is relat ively w ide sp rea d, tha t it var ies in occu rrence across different
segments of the economy, but that i t may not be a social force of over
w he lm ing em pir ica l im po rtanc e. Of cou rse, to the extent that var iat ion s in
credent ia l ism can be expla ined, they are w or th pu rsu ing in m ore deta i l .
As a s tep in this direction, Tables 9.6 and 9.7 presen t resu lts wh ich are
re levant to our pre l iminary hypotheses . There are three dependent var i
ables in these tables , one for each of the non-nested tests on which the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 209/375
92
cred ential m od el can be co m pa red to i ts r ival . Each variable is scored 0, if
the a lte rna t ive m ode l outperfo rm s the c redent ia l m ode l , if the resu lts a re
inconclus ive , and
2
if the creden tial m od el outpe rform s the al tern ative . T his
scor ing scheme is appropr ia te for the hypotheses under cons idera t ion ,
part icularly when the differing s lope and differing intercept models are
be ing com pa red , in so far as low scores are con sis tent with a h u m a n capital
in te rpre ta t ion and h igh scores wi th acreden t ia l l ing in te rpre ta t ion . H ow ev
r bec aus e the dis tr ib utio n of this variable is ra ther sk ew ed, OLS reg ression
is by pas sed in favor of a techn ique w hich is bet te r sui ted to this o rdin al
de pe nd en t var iable-ordered probi t ana lys is (Winship and M are , 1983) . The
coefficients have been est imated in such a way that posi t ive values mean
tha t an indus t ry wi th h igher va lues on the independent var iable a re more
likely to have higher credential scores. Table 9.6 contains the full range of
predic tor v a lues re levant to our hyp otheses , and Table 9 .7 prese nts t r im m ed
m od els w ith on ly those sets of variables w hich h av e significant pa rtial effects.
Star t ing wi th the hypotheses about the presence of b lack and female
w ork ers , there is l i tt le evide nce in favor of our ini tia l hy po thes is . W hile the
proport ion of black workers has an effect in the hypothesized direct ion, i t
ne ver app roa ch es s tat ist ical s ignificance. M oreo ver, there is qu ite s tro ng
evidence in favor of the proposit ion that industries with higher concentra
t ions of male, not female, workers tend to reward them in a credential is t
fashion.
Ev ide nc e in favor of the institu tion al effects is bes t de scrib ed as m ixe d.
Firs t , ha vin g a go ver nm enta l enti ty as a custo m er has l it t le to do w ith ho w
an ind us t r y respon ds to educa t iona l a tta inm ent . Second , w he n the indus t ry
is a go ve rnm en tal enti ty , there is a s ignificant tend enc y for edu ca tion to be
used in a nott-credential ized m ann er— a finding consis tent w ith Berg s
ob serv ation s in the early 70 s . Th ird, the coefficient for ex pe nd itu res on
adv er t i s ing a lso has the w ro ng s ign and
is
no t very large in any case. Fo urth ,
in those instances wh en service ind ustrie s differ system atical ly from othe r
gro up ing s , they , too , tend to reward educat ion in a con t inuou s , ra ther than
a s tep-w ise fashion. Fifth, ave rage firm size in an ind us try also a pp ea rs
unre la ted to c redent ia l i sm, cas t ing doubt on the bureaucra t ic r i tua l ism
hy po the sis . O n the other ha nd , ind ustries in the f inancial services sector are
s t rongly pred ispose d tow ard educat iona l c redent ia ll ing . In l ight of the
pat t e rn s for o ther service indu s t r ies , how eve r , i t
is
necessary to ask w het he r
this s tem s from this sector pr od uc ing an intangible pro du ct , or resu lts from
some other reason.
Th ere are also only we ak effects asso ciated w ith the experienc e v ariables .
Indu s t r ies wh ere exper ience cont inues to payoff wel l in to a wor ke r s career ,
i .e ., those with la te turn ing poi nts , are thos e wh ich eschew ste pw ise rew ard
functions in favor of continuous ones; that is they behave in a non-
cred entia list fashion. Th e sam e is true , alt ho ug h no t significantly so , for
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 210/375
193
TABLE 9.6 Between Industry Predictors of Educational Credentialism:
Ordered Probit Analysis. N=227.
ALTERNATIVE HYPOTHESES TO EDUCATION STEP
FUNCTION
Varying Slopes Low Cut Points High Cut Points
In tercep t
1
In tercep t
2
Experience
Turn ing Poin t
Tota l Retu rn
to
Experience
Female
Black
Coeff
-2.96
-1.82
-.032
1.71e 6
1.60
1.34
Requires Skill from
prev ious
job
Requires Skill
from School
Expend i tu re
on
Advert i s ing
Miss ing
Ad vert i s . Data
Average Fi rm
Size
Miss ing Fi rm
Size Data
Pet
of
Purchases
by G ov t
Miss ing G ov t
Data
Public
Admin i s t ra t i on
Finance
Service
Model
Chi -Square
-3.27
-1.15
-7.7e-4
.059
7.2e 3
.039
.016
.217
1.39
2.50
-.094
59.26
Chi Sq
18.79*
7.39*
3.69t
2.27
9.41*
.484
3.53
2.07
3.72
.030
1.49
.042
2.45
.567
5.75*
15.62*
.109
(15)
Coeff
-2.88
-1.31
-.020
6.0e-7
-.319
1.59
-3.41
-.375
-9.9e-4
-.378
8.3e-3
-.251
-.017
.169
.104
1.18
-.579
37.31
Chi Sq
17.92*
3.93*
1.52
.284
.368
.642
3.90*
.223
7.70*
1.15
1.70
1.69
2.77+
.342
.040
3.38
3.97*
(15)
Coeff
-2.39
-.900
-.020
-8.2e-7
-1.11
2.26
-3.99
1.07
-2.7e-4
.525
6.9e-4
.212
.010
.303
-.375
2.11
-.529
34.18
Chi Sq
13.38*
1.98
1.75
.550
4.68*
1.41
5.61*
1.90
.613
2.40
.014
1.30
1.15
1.19
.569
9.51*
3.56
(15)
Chi-Square significant,
p
<.05
Chi-Square significant, p <.10
indus tr ies wi th high tota l re turn s
to
exper ience . T hus ,
the
specula t ion tha t
credent ia ls
and
exper ience
are
a l ternative cr i ter ia
for
a l locat ing employ
m e n t r e w a r d s
is
t en ta tive ly sup por ted
in
this ana lysis .
P e r ha ps
the
m ost in teresting results
in
these tables , though, relate
to
w heth er jobs
in the
indus try require skil ls wh ich
are
ob taine d from formal
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 211/375
194
TABLE 9.7 Between Industry Predictors
of
Educational Credentialism:
Trimmed Ordered Probit Analysis. N=227.
ALTERNATIVE HYPOTHESES TO EDUCATION STEP
FUNCTION
Varying Slopes Low Cut Points High Cu t Points
Coeff
Chi Sq
Coeff
Chi Sq
Coeff
Chi Sq
Intercept 1
-3.08
21.98
Intercept 2 -1.96 9.25
Experience
Turning Point
-.031
3.76t
Total Return to
Experience -1.57e-6
2.52
Female
-1.57 9.29
Requires Skill from
previous job
-3.33
3.95*
Requires Skill
from School -1.42 4.78
Expenditure
on
Advertising -7.0e-4 3.30+
Missing Advertis.
Data
.124 .403
Public
Administration
-.968
3.49+
Finance
2.47
16.82
Service
Model
Chi-Square 54.58
(9)
-2.25 88.95
-.731 13.68
-2.73
-.288
-8.5e-4
-.389
2.99+
.219
6.23
4.01
-1.66
-.203
-.881
-3.83
.501
46.42*
.822
4.59
5.98
.625
.814
1.90
-.838 12.75
27.82
(6)
1.73
-.507
25.73
7.37
4.84
(5)
t
significant,
p
<.05
+1 significant,
p
<.10
tra ining in prev iou s jobs or require skil ls which are taug ht in
school.
I n b o t h
ins tances ,
and
p ar t icular ly
in
the case
of
sk il ls obta ined th rou gh formal
on-
the-job training , there are neg ative relat io nship s with the us e
of
educa t ion
as
a
c redent ia l .
In
o the r w ord s ,
in
those jobs w he re skills obta ined th ro ug h
formal training
are
impo r tan t , educa tion tends
to be
r e w a r d e d
in a
rela
t ively continuous fashion,
not in a
s tep-w ise fashion. W hat
is
impor tan t
about th is resul t , however , is tha t these data sh ow that both tendenc ies
opera te s imul taneous ly
in the
econom y. Cer ta in indus tr ies
fit the
h u m a n
capi ta l account of w hy schooling m atters: their jobs requ ire skil ls ob taine d
from formal training
and
they reward years
of
educa t ion
in a
relatively
s m oo th
and
con t inuo us fashion. Other indus tr ies
fit the
opp os i te p a t tern:
they
are
less lik ely
to be
concerned abo ut ski lls wh ich
are
associated w ith
formal learning experiences
and
they rew ard e duca t ion
in
discrete bloc ks.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 212/375
95
onclusions
This pap er s ta r ted w i th the prem ise tha t educ at ional c redent ia l i sm, as an
example of cul tura l symbol ism, impl ied the necess i ty of examining the
characteris t ics of those social actors , i .e . employers , who are the audience
for the sym bols in que st ion. Rathe r than postu lat in g a con stant cultu ral
tendency to a t tend to the symbol ic s ide of educat ional achievement , I
hypothes ized tha t d i f ferent par ts of the economic s t ruc ture would show
grea ter an d lesser proclivi t ies in this direct ion. In ad dit io n, extan t th eories
of labor m arke t organiza t ion we re used in a pre l im inary a t tem pt to expla in
pa t t e rns in be tw een- in dus t ry re sponse to educa t iona l ach ievemen ts .
The resul ts we re s t rongly supp ort iv e of the argu m en t s bas ic prem ise .
There are m easu rable an d cons is tent di fferences be tw een typ es of em plo y
ers in ho w they respo nd to var ious leve ls of educ at ional achiev em ent , a nd
these differences conform to the defini t ion of credential l ing derived from
ear l ie r work. Whi le the explanat ion of these be tween- indus t ry pa t te rns i s
pe rh ap s in a m ore form at ive s ta te , the out lines of an em erge nt explan at ion
can be de tec ted .
In genera l , hypotheses der ived f rom the ins t i tu t ional paradigm are not
cons is tent w i th these da ta . Ins tead, indus t r ies which u se edu cat ion in a
credential- l ike m an ne r are also those in w hich skills ob taine d from formal
train ing are no t req uir ed . These results also revealed a ten de ncy for
creden t ia l i sm to ho ld swa y in indus t r ia l se tt ings num erica l ly do m ina ted by
m en an d in the f inancial services sector. These pa tter ns sug gest there m ay
be a s t rong e lem ent of t radi t ional ism in educa t ional c redent ia l i sm. The
financial services sector remains among the most cultural ly conservative
groupings in our society, and with recent increases in the proport ion of
females in the labor m arke t , indus t r ies wi th h ig her pro por t ions of m en seem
ap pro pri atel y charac terized as m ore tradit ional . If this
is
the case, one m igh t
predict that even the l imited extent of educational credential l ing found in
this analysis , and in several othe rs , wil l bec om e scarcer as t im e goes by .
Notes
I am grateful for the assistance of
J
Beggs in helping with access to machine-
readable data sources; to L. Biggs for advice on certain computations; and to the
members of the U1C Social Organization thesis seminar for general comments.
Helpful sug gestions we re also prov ided by M. H ou t,
C
Bid well, and
S
Spilerman.
1. These theories differ s to wheth er the latent produc tivity tr its in question are
achieved in school or are inheren t in indiv idua ls. From the point of view of the
present paper, this distinction is not particularly relevant.
2.
Bidwell sees m atters differently, arguing that signalling theory s propensity to
construe education as a symbol means that it has an inherent tendency to posit
discon tinuities in the function which relates success to schooling. (127)
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 213/375
96
3.
His evidence for this prop osi t ion is that in a sam ple of Cal ifornia em ploy ers
those who had made the greatest efforts at racial integrat ion were those with the
h ighes t educat iona l requi rem ents . H ow th is should be in te rpre ted i s no t im m edi-
ately obv ious. Is the role of high edu cat ion al requ irem ents akin to sta tus
insurance for the organiza t ion 's publ ic image whereby members of cu l tu ra l ly
dev alue d group s can be safe ly incorpora ted on ly if they meet s t r inge nt s tand ards ?
Or, is the point that organizat ions for their internal control funct ions require a
certain am ou nt of cul tural exclusivi ty in their hiring pract ices w hic h wil l be
sat isfied one way or another?
4. This is su pp lem ent ed in their analysis with consid erat ion of the role of college
qua l i ty and th e role of majoring in a business subject . Spi lerm an an d L un de d o find
evidence that at tendance at a more select ive col lege and pursui t of a major in
bus ines s both act as crede nt ials , in the pro duc t ivi ty signal l ing sens e of the w or d.
Unfortunately, measures of these factors are absent in the data which we use.
5. He ob serve s this to be t rue in the priv ate sector, bu t not , ironical ly , in t he pu bl ic
sector.
6 . Co nsider , for exam ple, the repo rted correlat ion of .49 betw een F aia 's pri -
m ar y /s ec on da ry years o f educat ion var iab le and hav ing h igh school g ra du at io n
as one 's h ighes t educat iona l c redentia l . As is wel l know n, the m ax im um corre la-
t ions for variables with restricted rang es, a dich otom y l ike high school g rad ua t ion
is an extreme example, tend to be less than one, with the degree of at tenuat ion
depending on the correspondence in the marg ina l d i s t r ibu t ion of the var iab les
under examinat ion .
As an experim ent , I calculated the max im um possible correlat ion betw een Faia 's
p r im ary / sec on da r y educa t ion va r iab le and the h igh schoo l g rad ua t ion va r i -
able on the assu m ptio n tha t al l those, and only those, with 1 2,13 ,14 or 15 yea rs of
schooling had a high school degree as their highest level of certification. (Faia's
reported departures from perfect correlat ion are supposed to reflect the fact that
som e peop le finish hig h school w ith less than 2 years in school , an d tha t so m e w ith
12 to 15 years of school do not hav e a high school diplom a as their highest deg ree—
either beca use they failed to finish hig h school or because they ha ve alre ady finished
college) . The pr im ary /sec on da ry var iab le was coded in the sam e man ner as Fa ia ' s ,
i .e . as a spl ine variable with a m ini m um of 0 and a m ax im um of
12 .
The d is t r ibu t ion
of educat ion was taken from the 1980 PUMS fi le used in the subsequent analysis .
The correlat ion coefficient betw een prim ary /se co nd ary schooling an d high school
gra du at io n unde r these condit ions of perfect associat ion is .525. In othe r w or ds ,
Faia 's correlat ion is
93 %
of i ts m axim um value assum ing com ple te assoc ia t ion .
7.
Both tests were conducted on a subset of industries and the conclusions
reached on the basis of each test were identical .
8 . Prob ably, we unde res t im ate such effects since the data prob ably in clud e m or e
peo ple w ith too m an y years of educ at ion for their deg ree status tha n vice versa . If
so, these people wil l have negat ive residuals in the credent ial model , but wil l
pe rh ap s fall c loser to the regression l ine in the spl ine m odels .
9 . Tw o indus tries w ith less than 100 observ at ions, N ot specified professional
equ ipm ent ma nufac tur ing and Not spec if ied e lec tr ica l and ha rd w ar e prod ucts
wh olesa l ing w ere dro ppe d from the ana lys is . The ana lys i s reported here does
include data from 6 indus tries with betwe en 284 and 1000 incu m ben ts, and ano ther
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 214/375
297
30 industries w ith be tween 1000 and 1500 incumbents.
10. The two m easures were calculated as follows:
The total return measure was the area under the income - experience curve
(ignoring the portion constant across industries) calculated at 50 yea rs of experience.
The specific calculation w as:
TotRet = r | x
2
+ | x
3
|
The turning po int measure w as defined as:
TurnPt
=
- ~ -
2c
whe re b=coefficient for experience
c=coefficient for experience squared
11. To better probe the signalling hypothesis, it would be desirable to have
additional data. For example, industries in which employers agreed with the
statement, It is difficult to tell ahead of time who will work out on this
job
should
be those w hich rely on credentials as a signal. Unfortunately, these measu res are not
available for a broad cross-section of industries.
12.
Within
the
manufacturing sector, the transportation and utility sector and the
service sector, missing values were replaced by the within-sector m eans. In the
who lesale and retail trade sectors in which detailed industries were not covered in
the Input-O utput table, missing values were replaced by the overall mean. Again,
a dum my variable wa s included
in the
regressions
to
index those
cases
where means
were substituted.
13. Here, too, means were substituted for missing cases and a missing case
dum m y variable wa s entered as a control variable.
14. The basic data are those used by
Beggs 1994)
and were originally coded from
the 1977 and 1982 economic censuses. For industries where the economic census
data were unavailable, values were estimated using the following p rocedu re: In the
Annual D emographic File of the Current Population Survey, respond ents report the
num ber of peop le working for their employer. These estimates were aggregated to
the industrial level and for those cases where both the economic census and CPS
aggregates were available, the former were regressed on the latter. The resulting
prediction eq uation was then used
to
estimate missing
cases on
the economic census
measu re from the aggregated CPS figures. A dum my variable was created to
indicate which cases had been estimated in this way and it was also entered into the
final analyses.
eferences
Althauser, Robert, and Arne Kalleberg.
1981.
Firms,
Occupations, and the Structure
of Labor Markets: A Conceptual Analysis. In
Sociological Perspectives
on
Labor
Markets.
ed. Ivar Berg. New York: Academic Press.
Baron, James N., P. Devereaux-Jennings, and Frank R. Dobbin. 1988. Mission
Control? The Development of Personnel Systems in U.S. Industry.
merican
Sociological Review
53:497-514.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 215/375
198
Becker, G. 1964.
Human Capital.
New York: Colum bia U nivers i ty Press .
Beggs, John J. 1994. The Effects of the Institutional Environment on Inequality.
U np ubli she d m anu scrip t . Dep artm ent of Sociology. Louisiana State Universi ty .
Berg, Ivar.
1971.
Education
and Jobs: T he
Great
Training Robbery.
B oston: Beacon Pre ss.
Bidw ell , Ch arles E. 1989. Th e M eanin g of Educa t ional Attain m ent . Research in the
Sociology of E ducation and
Socialization 8:117-38.
Bielby, W il liam T,, an d James N . Baron. 1986. M en and W om en at Wo rk: Sex
segregat ion and stat is t ical discriminat ion. American
Journal
of Sociology 91:759-
99.
Bo urdieu , Pierre . 1977. Cu ltural R epro duct ion and Social Re prod uct ion . In
Power
and Ideology in E ducation.
ed. J . Karabel and A. H. Halsey. New York: Oxford
Universi ty Press.
Bridges , W i l liam P. , and W ayne
J
Villemez. 1991. Employment Relat ions and the
Labor Market : Integrat ing Inst i tut ional and Market Perspect ives. American
Sociological Revieiv 56:748-64.
Collins, Randall . 1979.
Th e Credential Society.
New York: Aca demic Press .
Cox, D . R.
1961.
Tests of Sepa rate Fam il ies of Hy poth ese s. In Proceedings
of the Fourth
erkeley Symposium on Mathem atical Statistics and
Pro bability.
ed. J. N ey m an . Vol .
1.
Berkeley: University of California Press.
D avids on, Russel l, and James G. M acKinnon . 1983. Some No n-nested Hy pothe sis
Tests and the Relations A m ong Th em .
Revieiv of
Economic
Studies
49:551-65.
DiM aggio, Paul , and John M ohr. 1985. Cu ltural Cap ital , Edu cat ional A ttainm ent
and Marital Selection.
American Journal of Sociology
90:1231-61.
DiM aggio, Paul J ., an d W alter W. Pow ell . 1983. The Iron Cage Revisi ted: Inst i tu
t ional Isom orp hism and Collective Rat ional i ty in O rgan izat iona l Fields. Ameri-
ca n Sociological Review 48:147-60.
DiPrete , Thomas. 1989.
The Bureaucratic
Labor
Market: Th e Case of the
Federal
Civil
Service.
N ew York: Plenu m Press .
Do bbin, Frank, John R. Sut ton, John W . Mey er, and W . Richard Scot t. 1993. Equ al
Opportun i ty Law and the Const ruc t ion of In te rna l Labor Markets . American
Journal of
Sociology 99:396-427.
Edelm an, Lauren . 1990. Legal Envi ron men ts and Org aniza t iona l Go vern ance .
American Journal of Sociology 95:1401-40.
Faia, M ichae l. 1981. Selection by Certification.
American Journal of Sociology
86:1093-1111.
Fe athe rm an, D avid L., and Robert M. H aus er. 1978.
Opportunity and Change.
N e w
York: Academic Press.
Halaby , Char les N ., and D avid L. W eakl iem,
1993.
Ow nersh ip and Au thori ty in the
Earnings Funct ion: Non-nested Tests of Alternat ive Specificat ions.
American
Sociological
Review
58:16-30.
Mey er, John, and Brian Row an. 1977. Inst i tut ional ized O rgan izat ion s: Form al
St ruc ture as Myth and Ceremony.
American
Journal
o f
Sociology 83:364-85.
Pesa ran, M. H. 1974. O n the Ge neral Problem of M odel Select ion. Review of
Economic S tudies 41:153-71.
Spenc e, Michael A . 1974.
Market Signaling.
Cam bridge , Mass . : H arva rd U niv . Press .
Spi le rman, Sey mo ur an d Tormo d Lund e . 1991 . Fea tures of Educat ional A t ta in
ment and Job Promot ion Prospec ts . American Journal of Sociology. 97:689-720.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 216/375
199
Stiglitz, J. 1975. The Theory of 'Screening', E ducation, and the Distribution of
Income.
merican Economic Review
65:283-300.
Weakliem, David L. 1992. C omparing Non-nested Models for Contingency Tables.
In
Sociological Me thodology 1992.
ed. Peter Marsden. Vol. 22. Oxford: Basil
Blackwell.
Williamson, Oliver. 1975.
Markets and Hierarchies.
New York: Free Press.
W inship, Christopher, and Robert M are. 1984. Regression Models w ith Ordinal
Variables. merican Sociological Review 49:512-25.
Zucker, Lynne G
1987.
Institutional Theories of Organ izations.
nnual Review of
Sociology
13:443-64.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 217/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 218/375
1
Education, Earnings Gain, and
Earnings Loss in Loosely and Tightly
Structured Labor Markets:
A Com parison betw een
the United States and Germany
Thomas
A.
iPrete
Duke University
Patricia A McManus
Indiana University
The Link between Education and Earnings
Education is positively associated with earnings in all industrialized
coun tries. The hu m an capital explanation for this connection is that
education pro vides skills, that skills re costly to obtain, and that the market
therefore pays a re turn to these skills. Sociologists and institutiona l econo
mists recognize the connection between ed ucation and earnings
to
be more
complex than this skill-focused explanation suggests, for several well-kno wn
reasons. W hereas at least some skills are comm on to all countries at a given
level of technological development, the way these skills are combined in the
education process tends to show more cross-national variation. Further
mo re, the credentials prov ided by the educational system show substantial
cross-national variation, and the process by which individuals convert
skills and creden tials into job incum bency also differs by country . Finally,
earnings are a function not merely (some would argue not even primarily)
of on e s skills, bu t also of one s occupation, and particularly of one s job. In
short, the connection between earnings and skills is mediated by the
processes that link earnings with jobs and occupations and that link peop le
with jobs.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 219/375
2 2
Much of the early sociological study of the linking processes between
education and occupation involved m easu ring the effect of education on the
attainm ent of occupational status. More recently, it has involved the stud y
of how firms and corporate occupational bodies link or do not link training
with the educational system. Sociology has also recently been concerned
with the related question of how educational and training certification
struc ture opportu nities for occupational access and occu pational mo bility.
Differences in the process by which indiv iduals get access to occupations
and change occupations is of course w hat scholars m ean (at least in part) by
the term labor market structure. Thus, the new literature goes beyond either
the hum an capital approach to education (education —
skills — learning s)
or the status attainment tradition (education — ^occupational attainment) in
focusing on the specific mechanisms by which individuals obtain and
change jobs. These new studies in effect address the question: ho w does
labor market structure create a linkage between education and labor m arket
outcomes?
One important finding of the recent literature concerns the level of
linkage between education and w ork institutions. Scholars hav e arg ued
that education-work linkages are relatively tight in many continental
countries and also in Japan (though the structure is quite different), wh ile
the linkages are relatively loose in the United States (e.g., Rosenbaum et. al.
1990).
T his assertion can be taken in two different
senses.
First, the process
of matching individuals and jobs is relatively unsystematic in the United
States. W orkers jum p from job to job at the start of their career learn ing
abou t and evaluating their opp ortun ities before they finally find a
jo
that
is suitable to their abilities and skills. In contrast, the school-to-work
transition is more orderly in many other indus trialized countries: the more
standardized credentials make clear which occupation a particular indi
vidua l is trained for, and the process of acquiring a suitable
jo
thus occurs
relatively quickly. Second, the statem ent that school to wo rk linkages are
weak in the United States can be taken to mean that even after the more
lengthy Am erican early career sorting process is conc luded, the outcom es
associated with a particular level of educational attainment or with a
particular educational certificate or degree are more variable in the U.S.
This variability exists because the m ore loosely structured Am erican system
does not preven t workers from obtaining jobs for which their education is
not an obvious match to the extent found in mo re rigid systems.
The early career implications of tight or loose linkages betw een educa
tion and work have been thoroughly researched (Rosenbaum et al. 1990).
Empirical investigations of how the school-to-work linkage affects the
relationship between education and subsequent career experience, how
ever, are less complete. Yet, the prediction abou t this relationship that
follows from the above characterization of the school-to-work linkage is
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 220/375
2 3
rather ap paren t. If linkages betwee n education and job are tight in the early
career, then one would expect education's subsequent impact on career
processes to be mediated by initial placem ent. Subsequent outcom es
should largely be a function of one's occupational and job position. This is
particularly true if more structured labor markets create tighter links
between occupation or job and rewards, as some scholars have found
(Haller 1987). In contrast, individual resources should continue to affect
career outcomes in more loosely structured labor markets, even when
positional resources are controlled. The contrast between countries w ith
weakly and strongly linked institutions should be even stronger in rela
tively turbulent economic periods— such as we are currently experiencing
in the industrial world—because individual resources become still more
imp ortant in loosely structured, turbulent labor ma rkets. In other w ord s,
the looser linkages between education and early job placement
in
the United
States shou ld lead to stronger direct linkages betw een edu cation and caree r
dynamics in the United States than in countries where labor markets (and
particularly the school to work transition) are more structured.
This paper explores this inference by analyzing edu cation's
role in
career
earn ings change in two countries: Germany and the U.S. Differences
between the Germ an and Am erican educational systems and their linkages
to the world of work h ave been extensively discussed in the literature, and
need not be rep eated in full here {e.g., Konig and Muller
1986,
M ayer et al.
1989; Blossfeld et al. 1989). Two differences need to be emp hasized. First,
the connections between education or training and the world of work are
m ore systematic in Germ any tha n in the U.S., because the Germ an system
produ ces stand ardiz ed education and training credentials that are a prereq
uisite for most skilled
jobs.
Second, job-specific training is occupationally
based in Germ any, while in the U.S. a great deal of job-specific train ing is
pro vide d w ithin firm-based internal labor mark ets as well as within occu
pations: the American system generates earnings trajectories that tend to
rise both with years of tenure in a firm and with years of tenure in an
occupation (DiPrete and McManus 1993).
Differences between
the
German and A merican labor ma rkets lead to the
following specific hyp othese s. [1] We expect the direct effect of edu cation
to be stronger in the United States, because individual resources matter
more in a society wh ere the institutional glue that collectivizes life
chances is stronger.
[ ]
W e expect the indirect effect of educ ation (through
occupation) to be stronger in Germany because of the tighter linkage
between education/training and work. [3] In contrast to Germ any, w here
training affects e arnings largely throu gh credentials and thro ugh the occu
pation to which these credentials provide
access,
the influence of training on
labor market outcom es is revealed in the U.S. as effects of ten ure in the firm
or tenure in the occupation. The cross-national difference thus co ntrasts the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 221/375
2 4
effects of incumbency in a par t icular s tatus and tenure in that s tatus .
While da ta l imi ta t ions prevent us f rom compar ing his tor ica l per iods ,
cross-na tional differences in these reg ard s should b e par t icu lar ly s tro ng in
the 1980s relat ive to ear l ier t imes because of the par t ial breakdown of
Amer ican internal labor markets tha t l ink individual outcomes together ,
and the consequent growing impor tance of individual level resources in
dete rm inin g labor m arket outcomes . The cross-nat ional contras t sho uld
also be especially s tron g in recessions for the sam e reason: du rin g recessions
the internal labor markets that collectivize outcomes become weaker , and
ind ividu al resources beco me a s t ronger predic tor of outcom es . O ur data do
allow us to evaluate this second prediction.
ata and Methods
German data are f rom the Engl ish- language vers ion of the German
Socio-E conom ic Pan el (GSOEP) an d cover the six-year pe rio d from 1985 to
1991. Given the recency of the German reunif ication, we restr ict our
atten tion to the W est G erm an sam ple , tho ug h w e take account of reunif ica
t ion in the interpretation of the West German results .
2
American data for
1985-1991 are taken from the Pane l Stu dy of Inco m e Dyn am ics (PSID). W e
divid ed this s ix year per io d into two three-year segm ents , an d analyze
three-y ear earn ing s mobili ty f rom 1985-1988 an d th en from 1988-1991. The
f irst per iod is one of relat ive prosp eri ty in both c oun tr ies , w hile th e seco nd
pe rio d is the s tar t of a recession in bo th cou ntr ies . The second p er io d also
encompasses the German reunif ica t ion.
Results in this pa pe r are co m pu ted for men on ly w ho w ere in on e of the
follow ing tw o categ ories: [1] 18-61 ye ars of age and em pl oy ed in 1985, an d
sample respondents in the years 1985-1988, or [2] 18-61 years of age and
em plo ye d in 1988, an d sam ple resp on de nts in the years 1988-1991. The
males in the German sample , when weighted, cons t i tu te a representa t ive
sample of the wes tern German popula t ion, inc luding the fore ign popula
t ion res ident in W est Ge rman y. The males in the PSID, w he n w eigh ted, a re
repre sen tative of Am erican male he ad s of ho us eh old s in 1968. Because
m uc h of our in teres t i s in pop ula t ion com par isons , we use w eigh ted PSID
samples for our cross-national comparisons of earnings mobili ty (heads
who are not descendants of or iginal 1968 PSID sample families are given
zero we igh ts in the
PSID).
In our multiv ar iate analy ses desig ned to interpret
cross-national differences, w e use the full (unw eighted ) PSID sam ple.
W e descr ibe e lsewhere (DiPrete and Mc M anu s 1995) the proc edu re w e
use d to est im ate earn ing s mo bili ty. Brief ly, we con structed a m ea su rem en t
m od el for earnin gs in G erm any an d in the U.S. , in ord er to remo ve from
measured earnings mobi l i ty tha t por t ion which is due to measurement
error in the earnings reports at the two points in t ime between which
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 222/375
205
mobility is m easured . The Germ an latent earnings variable
is
based on the
following two measures:
[ ] GEARN1:
The respo nde nt's answer to the question: How high were
your [gross] earnings last m on th? It includes overtime but not holiday or
back pay.
[2] GEARN2:
The retrospective report from the following year survey
of
average gross monthly earnings for months in which the respondent
received earnings . Respondents are asked to report earnings including
sick-pay and compensation payments (but excluding bonuses such as
holiday money), payments for training, self-employed earnings and earn
ings from second jobs.
We then cons tructed a similar measurem ent m odel for the PSID sam ple
based on the following two m easures:
[ ] PEARN1:
The total labor earnings
in a
calendar year
as
reported in the
following PSID survey , divided by 12.
[2] PEARN 2: The projected ma in job earnings for a calendar year com
puted from the wage per hour on the main
job
as reported in that year, the
num ber of weeks w orked on main jobs that year
as
reported in the following
yea r's survey, and the average hours p er week on ma in jobs in the calendar
year as reported in the following year's survey.
We corrected the earnings data in both coun tries for inflation. We used
CPI-U (The U.S. City Average for all Urban Consum ers) for the US, and
r is
und Preisindizesfilrdie ebensha ltung
for Germany. In both the German
and
the
American measurem ent
model, we
specified the metric of the latent
earnings variable to be the same as the first of the two earnings measures
(GEARN1 or PEARN 1).
Results
We describe elsew here (DiPrete and M cManus 1995) the major differ
ences in earn ings mobility
in
the two countries. Briefly stated, Germ any had
substantially greater earnings growth during this period than did the
United States. Furtherm ore, Germany h ad less variance in earnings out
comes. While part of the greater turbulence in the Am erican resu lts is due
to the greater levels of job mobility found in the
U.S.,
much of the earnings
dynam ics occurs within occupations and w ithin
jobs
in
the
same firm. Here
we focus on the effects of education and training on these earnings dynamics.
As a baseline against w hich to interpret later analyses, we first com pare
the relationship between education level and the log of 1985 monthly
earnings for Germany and the United States. In this baseline model, we
specified earn ings to be a function of age, education and training.
3
Because
the educational systems of the two countries differ, we used different
educational benchm arks in the Germ an and Am erican analyses. For the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 223/375
206
United States , we made the following mutually exclusive dis t inctions: (1)
less than high school, 2) high school com pletion , (3) college atte nd an ce, a nd
(4) col lege comp let ion and pos t-gra dua te educat ion . In G erm any , we
similar ly dis t ing uish ed four m utu ally exclusive edu cation al levels (1) thos e
w ho had no t comple ted G erma n lower schooling Hau ptschule, Realschule),
inc lud ing those w ith no degree , foreigners w ith a deg ree belo w th e college
level , an d G erm ans w h o repor t ano ther degre e below the univ ers i ty level
Anderer Ab schlufi)
other tha n those inclu ded in one of the categories be low ;
(2 ) t h o s e w h o h a d c o m p l e t e d l o w e r i n t e r m e d i a t e s c h o o l i n g
Hauptschulabschlufi, Mittlere Reife, Realschulabschlufi),
(3) those w h o ha d
completed upper - intermedia te school ing
Fachh ochschidreife- Abschlufi einer
Fachoberschule
or an
Abitur - Hochschulreife),
and (4) those w ho ha d com
pleted technical school or universi ty Hochschulabschlu.fi ,
Fachhochschule,
Universitat, Technische/Sonstige Hochschule
or un iversi ty d egre e from out
s ide Ge rma ny) . Educ at ional level i s m eas ure d eve ry year in the GSOEP. In
the P SID, the educa tional level of all hou seho ld h ead s was m easu red in 1985.
In add i t ion, w e m easu red add i t ional wo rk- re la ted t ra ining. For Ger
m an y w e used the fol lowing three ca tegories : (1) those wh o had com pleted
an app rentice ship or vocational school Lehre,
B erufsfachschule, Han delsschule,
Schule des Gesundh eitswesens), (2) those who completed Fachschule, and (3)
those with other vocational training
Sonstige Ausbildung).
The Uni ted
States does not have the extensive vocational training system found in
G erm an y, bu t the PSID includes the following ques t ion: Did you receive
any other degree or a cer t if icate through a vocational school, a training
school , or an apprent icesh ip pro gra m ? W e include d the ans we r to th is
question as a var iable in our analyses.
W e nex t specified a m ore elaborate regression that includ ed th e baseline
variab les as well as var iables wh ich mig ht m edia te the effects of ed uca tion ,
namely occupat ion, indus try , tenure wi th cur rent employer , and ( in Ger
many) whether the worker had a te rm or an indef ini te employment con
tract.
4
Given the re la tive ly sma l l samp les found in panel da ta , w e l imited
ou r occu pation al controls to mo dif ied EG P categories (Ga nzeb oom , Luijkx,
and Tre im an 1989), and use d sem i/un ski l led m an ua l w ork as the om it ted
category. Indu s try w as m easu red us ing a com m on four-sec tor taxon om y,
which we fur ther e labora ted for the manufactur ing/mining sec tor .
5
Table 10.1 sho w s the results f rom the baseline earnin gs m od el. These
resul ts dem ons tra te tha t educat ion has a powerful influence on earnin gs in
bo th the U.S . an d Ger m an y. The effec ts of col lege co m ple t ion vs .
in termedia te- level school ing were very s imilar in the two countr ies
(0.548-0.173 = 0 375 in Ge rm an y, a nd 0.784-0.372= 0.412 in the U nit ed
States).
Two differences between the U.S. an d Germ an y are nota ble. Firs t ,
t ra ining ha d a big impact on earnin gs in Ge rman y. A w orke r wi th
lower- intermedia te educat ion
Hauptschulabschlufi, Realabschlufi
wh o c om -
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 224/375
2 7
TABLE
10 1
Effects of Education and Training on Log Earnings Germany
and the United States,
1985,
Males only
Variab le
Intercept
Age
A ge
2
Lower School
Middle School
Univers i ty
Vocat ional Tra in ing
Fachschule
R
2
Intercept
Age
A ge
2
High School
Some College
College Degree
Extra Certification
R
2
G e r m a n y
Coefficient
4.939
.068
-.001
.127
.173
.548
,095
.247
.28
United States
9.833
.083
-.001
.372
.488
.784
.019
.22
T Ratio
59.252
15.465
-13.653
8.594
5.105
25.483
6.904
9.568
60.119
9.428
-7.574
10.770
13.209
22.478
.779
pleted both an apprenticeship and Fachschule (and most of those with
Fachschule had only completed Hauptschule or Realschule had the same
expected earnings as did a respondent w ho had graduated from technical
college or un ive rsi ty (0.173+0.095+0.247= 0.515 vs . 0.548). The ready
availability of economically valuable training credentials to workers who
did not a ttend college implies narrower ea rnings differentials across levels
of formal education than one finds in the United States, where extra
certification ou tside of school had no obvious economic
value.
The second
distinctive cross-national difference evident in table 10.1 concerns the
earnings of Americans who did not complete high school. The earnings
pena lty for this group is considerably larger than for the comparable g rou p
in Germany.
Table 10.2 then elaborates table
10 1
by including the intervening vari
ables mentioned above. These results suppo rt the claim that educa tion and
job placement are more tightly linked in Germany than in the U.S. and
further suggest that the consequence of
a
looser linkage between education
and job implies a stronge r direct effect of education on earn ings in the
United States than in Germany. In the United States, the gap between the
effects of college completion and high school shrank only slightly when
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 225/375
2 8
TABLE 10.2 Log Earnings, 1985 for the U.S. and Germany
United States Germ any
Estimate
T Ratio
Estimate
T Ratio
Intercept
Age in 1985
A ge
2
High School
Som e Col lege
College Degree
Other Cert .
H a u p t / R e a l
Abi tur
U n i / F a c h h o c h
Apprent icesh ip
Fachschule
U p p e r P r o f / M g r
Lower Pro f /Mgr
Rou t ine Nonman
Self-employed
Manua l Supe rv
Ski l led manual
Empl . Tenure
Un ion Member
Light Mfg
C o m m / T r a n / U t i l
W h o l / R e t T r a d e
Services
Term Cont rac t
R
2
10.320
0.061
-0.001
0 228
0 322
0 508
0.018
—
—
—
—
—
0.621
0.421
0.180
-0.070
0 363
0 250
0.017
0.181
-0.200
0 005
-0.200
-0.236
.3
66.927
7 438
-6.735
7.041
8 978
13.146
0 786
—
—
—
—
—
13.505
12.671
4.384
-1.299
7 000
8 072
10.382
6 876
-6.082
-0.129
-5.497
-7.440
5 380
0 048
-0.001
—
—
—
—
0 036
0 004
0.191
0 070
0.144
0 550
0 227
0.157
0 394
0.137
0 062
0.001
0 034
-0.054
-0.110
-0.138
-0.173
0.111
.4:
65.269
11.161
-10.229
—
—
—
—
2.252
0.121
5 398
5 038
5 667
21.994
10.625
5 493
11.278
4.611
3 727
1.202
2.930
-3.690
-4.967
-5.277
-9.914
-4.237
intervening var iables were included (from .412 to 0.28) , while the gap
betw een univers i ty and lowe r- intermedia te school ing in G erm any shra nk
m or e subs tan tially (from .375 to 0.155). S imilarly, the con tracti on is grea ter
in G erm any for the contras t be tw een univ ers i ty comp let ion and the low est
(omit ted) educ at ional ca tegory. O ne should note , how ever , tha t im por ta nt
direct effects of edu catio n an d training rem ain for G erm an m ales even after
job attr ib ute s are take n into accoun t. Th e existence of m od era tely s tro ng
direct effects of education and training suggest that job rewards differ in
G erm an y even w ithin the occupat ional and ind us tr ia l ca tegor ies specified
in table 10.2, an d th at this w ithin -categ ory varia tion is directly link ed to
ed uca tion . To ph rase i t dif ferently, edu catio n an d trainin g creden tials
con tinu e to affect e arnin gs levels in G erm an y ev en after gro ss occu pation al
and industry dis t inctions are taken into account.
The contrast between the s trong effects of training on earnings in Ger-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 226/375
2 9
m an y a nd the absence of t ra inin g effects in the Un ited S tates is cou nterb al
anced by the con tras tb e tw een the v i r tua l absence of em ploy er tenu re e ffec ts
in G erm any a n d the i r s t rong influence in the Un i ted Sta tes . In hu m an
capi ta l m od els of Am erican e arnings , emplo yer tenu re i s typica l ly taken as
an indicator of f irm-specific t ra ining. From a h u m an ca pital pe rsp ecti ve ,
one mig ht in te rpre t the di ffe rent pa t te rn in the tw o coun tr ies as imp lying
that the tw o system s for teaching job-related skills are functionally equ iva
lent: Am erican wo rke rs a re rew ard ed for informal t ra in ing, wh i le G erm an
w ork ers a re rew ard ed for t ra in ing credent ia ls . H ow eve r , th is in te rpre ta
t ion w ou ld be inad equ ate for two reasons . Fi rs t, the absence of em plo yer
ten ure effects on ea rni ng s in G erm an y rem ains even if t ra inin g variables are
removed f rom the mode l .
6
Second, the benefi ts of employer tenure in the
U.S.
are s imilar a t a l l levels of edu cation al a t ta in m ent. In other wo rd s, the
ave rag e Am erica n wo rke r wh o lacks a college edu catio n doe s not offset this
dis ad va nta ge via em plo ye r training to s tate this m ore technically, there is
no pos i t ive in te rac t ion be tw een high school d ip lom a an d emp loye r exper i
ence) ,
whi le German workers who lack a col lege educat ion can c lose the
ea rn ings ga p be tw een them se lves and co l lege educa ted wo rke rs by ob ta in
ing training certification.
7
Earnings re tur ns to tenu re a re therefore a
dis t inct ive characteris t ic of Am erican labor mar kets ; they co rres po nd to bu
are no t equiv a lent to re tu rns to t ra in ing credent ia ls in Ge rm any . The
em plo yer te nu re e ffect in the U.S. i s un do ub ted ly an
average
effect th at
m ask s subs tant ia l var iance a t the individua l leve l. But w hi le an im po rtan t
component of such var ia t ion in Germany is manifes t ly measurable v ia
train ing c rede ntials , the varian ce in the U.S. is la tent: it is an im po rta nt
aspect of Am erica n labor ma rke t s tructu re, b ut i t is no t easily ch arac terize d.
Having es tabl ished the connect ion be tween educat ion and earnings in
the two countr ies , w e next focus on the re la t ionship be tw een e duc at ion an d
earn ings change . Table 10.3 exam ines the connect ion be tw een earn ings
dy nam ics and educat io n in the s imples t poss ib le w ay. Earning s cha nge is
repre sente d as a d ich otom ous v ar iable in wh ich earnings ga in ne t of
inflat ion) is dis t inguished from earnings decline.
8
The table shows a clear
re la t ionship be tw een earn ings ga in and educ at ion in the Un i ted Sta tes, and
a m ore am bigu ous , weake r re la t ionsh ip in Ge rma ny . Overa l l , up w ar d
mobi l i ty was much more common dur ing th i s pe r iod in Germany than i t
w as in the U ni ted Sta tes , and the f requency w i th which i t occurre d app ear s
to be largely ind ep en de nt of edu catio nal level . In the la te 1980s, ear nin gs
m obil i ty in G er m an y could accu rately be desc ribed as a collective eve nt that
contras ts wi th the la rge var iance in individual outcomes in the Uni ted
States . S om e of this varian ce in the Un ited States is ap pa ren tly at tr ib utab le
to edu cati on , s ince there is a noticeable patt ern for college gr ad ua tes to hav e
high er probabi l i ties of earnin gs ga in than less edu ca ted w ork ers . The
pat te rn i s s t ronges t when comparing the ext remes of the educat iona l
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 227/375
210
TABLE 10.3 Earnings Mobility
Sample Limited to Those Employed a t Start and End
ermany
1985-1988 1988-1991
Upward Downward Upward Downward
No ne/O ther 83 17 77 23
Lower Intermediate
No Apprenticeship 83 17 78 22
Appren/Vocational
Train ing 82 18 74 26
Fachschule 83 17 78 22
Up per-Interm ediate 88 12 71 29
Un iversity 83 17 71 29
Less than High School
High School Comple t ion
Some College
College Complet ion
United States
1985-1988
Upward Downward
58
60
60
66
42
40
40
34
1988-1991
Upward Downward
43
51
48
54
57
49
52
46
Sample Includes Those Not Working at Destination Time
ermany
1985-1988 1988-1991
Upward Downward Upward Downward
N one /O ther 72 28 71 29
Lower-Intermediate
No Apprenticeship 73 27 71 29
Appren/Vocational
Train ing 76 24 67 33
Fachschule 81 19 75 25
Up per-Interm ediate 83 17 62 38
University 79 21 68 32
Less than High School
High School Comple t ion
Som e Col lege
College Complet ion
United States
1985-1988
Upward Downward
52
56
56
63
48
44
44
37
1988-1991
Upward Downward
36
45
45
50
64
55
55
50
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 228/375
2
dis t r ibut ion: the od d s tha t a college gra du ate exper ienced a rea l earnin gs
gain over three years w ere 30 high er than a high school gra du ate d ur in g
1985-88 and 40 high er than a w orke r w ho ha d not com pleted high school .
The od ds ra t io for the contras t be tw een col lege grad uate s an d hig h school
grad ua te s na r row ed so me wh at dur in g the second pe r iod , bu t the gap
between col lege graduates and high school dropouts grew larger in the
second per iod than the f irs t .
Th e result s desc ribe d in table 10.3 are further con firm ed in table 10.4,
w hic h sho w s the resu lts of logist ic regression s on the proba bil i ty of earn ing s
gro w th betw een 1985 an d 1988 an d betw een 1988 an d 1991. W hen age and
hours of work in the base year are controlled, American workers with
college degrees were noticeably more l ikely to experience earnings gains
than were workers with less education in both the f irs t and in the second
per io d. T he G erm an contras t be tw een the top and bot tom of the ed ucat io nal
hiera rchy is nu m erically as large as the contrast in the Un ited States for th e
fir st per iod , tho ug h the s tan da rd er rors are la rger in G erm any . Con s is tent
w ith ou r predic tions , the effects for the U.S. an d G erm any m ov ed in
different direction s in resp on se to the less favorable econo mic co ndit io ns in
1988-1991.
The recess ion heigh tened the contras t be tw een the earning s of
Am er ican wo rke rs w ho ha d completed high school and those w ho had not .
In G erm an y, full recessio n d id no t hit un til 1992, bu t the pe rio d 1988-1991
w as one of econom ic s low do w n, exacerbated by the economic unce r ta inty
wh ich acco mp anied reunif ica tion. The s low do w n actual ly redu ced the
effect of education or the direction of earnings change in Germany.
Table 10.4 is bas ed only on the quali tat ive com paris on of w he th er real
earn ing s rose or fell over time. On e m ight ques tion w he the r the cross-nation al
contrast would be as great if the analysis focused on changes in earnings
ra ther than a dich otom ous m easu re of earning s increase . W e ad dres s th is
qu estio n by rep ortin g regressio ns of the log of the rat io of ea rnin gs cha ng e
in table 10.5.
9
The results without controls are presented in the lef t hand
columns of each panel, while the results with controls for occupation,
indu s try , emplo yer ten ure and other re levant variables are prese nted in the
r ight ha nd pa ne l for each set. B oth sets of results sugge st that cross-natio nal
com par ison s based on the m ag ni tu de of earning s chang e differ f rom
cross-nation al com par ison s base d on the direction of earn ing s cha nge . In
the first period (a period of prosperity), the total effects of education on
earning s we re ac tual ly grea ter in G erm any than in the Un i ted S ta tes , which
is no t wh at our analys is of the di rec t ion of chan ge show ed. Edu cat ion
ef fec ts on the magni tude of change, however , d isappeared in Germany
dur ing this per iod of economic s lowdown, whi le they increased in the
recessionary U .S. These results are consis tent with the results we obta ined
us ing the di rec t ion of earnin gs change as the de pe nd en t var iable . The
incre ase in the size of th e effects in the U nite d States can largely be att rib ute d
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 229/375
TABLE
10 4
Effect
of
Education
on
The Probability
of Upward Earnings Mobility
Intercept
Age
A ge
2
High School
Some Col lege
College Degree
Extra Certification
L o g h o u r s /
m o n t h in base year
Intercept
Age
A ge
2
L o w e r / I n t e r m e d i a t e
Upper - In t e rmed ia t e
Univers i ty
A p p r e n / V o c s c h o o l
Fachschule
New Tra in ing
stimate
8 082
-.08
.001
.069
.167
.365
-.024
-1.142
8 094
-.044
.0003
.17
.419
.342
-.254
.049
.372
United States
1985-1988
Standard
rror
1 0316
.0325
.0004
.1256
.1345
.1274
.0882
.165
1 969
.046
.0006
.14
.366
.216
.135
.252
.753
Prob
0 0001
0.0142
0.0788
0.581
0 2155
0 0041
0.7836
0 0001
stimate
5 5479
-0.0781
0.0007
0.2599
0.2912
0.4992
0.0214
-0.7656
ermany
0 0001
0 3361
.5685
0.2234
0.2514
0.1126
0.0596
0.8452
0 6215
3.511
-.047
.00033
-.074
-.329
-.134
- .137
.231
.126
1988-1991
Standa rd
rror
1 1265
.038
.0005
.1378
.1456
.1375
.0934
.1622
2.19
.052
.00063
.162
.328
.201
.14
.239
.656
Prob
.0001
.0396
.1289
.0594
.0455
.0003
.8188
.0001
0.1088
0.3638
.6003
0 6473
0 3155
0 5053
0.3279
0.3337
0 8475
to the exper ience of those who lack a h igh school educat ion who did bad ly
in the second per iod.
1 0
The recession also changed
the
pa t te rn
of
training effects
in
Ge r m a ny .
Vocational training had no discernible effect on ea rn ings change in e i ther
pe rio d. Th e effect of ga ining new t ra ining credent ia ls du r in g the three yea r
pe r iod was im por t a n t in the first p erio d but not the s econd whi le hav ing a
chschule c ertificate was im por t a n t in the second per io d but not the first.
Th e lack of stability in the Ge rma n resul ts m ay be due to the relat ively small
n u m b e r
of
wo r ke r s
who
ga ined
new
t raining credentials over
a
relatively
shor t per iod
of
t ime. Overall
the
results sugg est that training l ike ed uca
t ion
has an
impac t
in
G e r m a n y
on
ear nin gs trajector ies ap ar t f rom br oa d
occupa t iona l
and
indus tr ia l ca tegor ies thou gh
the
precise form
of
this
re la t ionship is s om e w ha t a m b iguous .
The results discussed so far are genera l ly but not comple te ly cons is tent
wi th the hypotheses p resen ted at the start of the p a pe r . In genera l the
effects
of
educa t ion
are
more media ted
in
Ge r m a ny tha n
in the U S
Howe ve r
the
cross-nation al con trast
is not as
cr isp
as it
cou ld be— educa-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 230/375
Intercept
Age
A ge
2
High School
Some Col lege
College Degree
Extra Certification
Log hou r s / m on t h i n ba s e ye a r
R
2
Intercept
Age
Age
2
Low-Intermedia te
Upper- In termedia te
Univers i ty
Appr e n / Voc s c hoo l
Fachschule
New Tra in ing
Log hou r s / m on t h i n ba s e ye a r
R
2
TABLE 10.5 Effect of Ed uca tion
1985-1988
Total Effects
stimate
2.156
-.017
.0002
-.005
.016
.031
-.01
-.33
.788
-.011
.00011
.0083
.035
.062
-.001
-.001
.1112
-.083
T Ratio
15.1
-3.3
2.4
-0.3
0.7
1.5
-0.7
-15.1
.09
6.7
-3.9
3.1
0.9
1.8
4.9
-0.1
0.1
3.1
-4.0
.05
on the Log-Ratio of Earnings
United States
Partial Effects
stimate
—
-.008
.0001
.018
.03
.057
-.005
-.259
—
-.004
.00004
.0045
.052
.089
.008
.009
.0709
-.042
T Ratio
—
1.5
1.0
0.9
1.3
2.3
0.4
-10.6
1988-1991
Total Effects
stimate
1.746
.003
-.0001
.056
.064
.092
.005
-.342
Germany
—
-1.4
1.0
0.4
2.4
5.0
0.9
0.6
2.0
-1.9
.518
.001
-.00004
-.004
-.022
-.015
-.001
.04
-.004
-.078
T Ratio
8.0
0.4
-1.2
2.0
2.2
3.3
0.2
-11.3
.07
3.1
0.2
-0.9
-0.3
-0.9
-1.0
-0.1
2.2
-0.1
-2.7
.03
Partial Effects
Estimate
—
.004
-.0001
.0644
.062
.087
.004
-.182
—
.005
-.0006
.01016
-.003
.029
.006
.062
.002
.009
T Ratio
—
0.5
-1.4
2.3
2.0
2.6
0.2
-5.2
—
1.1
-1 7
0.7
-0.1
1.4
0.5
3.4
0.03
0.3
ISJ
- J
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 231/375
4
TABLE 10.6 Age-Education Interaction Effects on Log Earnings, 1985
Variable
Intercept
Age
A ge
2
High School
Some College
College Degree
Extra Certification
H S * (Age-18)
Col lege
(Age-25)
College Degree
(Age-25)
Intercept
Age
A ge
2
Lower-In termedia te
Upper - In t e rmed ia t e
Universi ty
Vocat ional Training
Fachschule
Lower-In ter (Age-18)
Upper - In t e r
(Age-18)
Univers i ty
(Age-25)
United States
Estimate
10.397
.06
-.001
.213
.292
.452
.016
.00008
.002
.004
5 352
.053
-.001
.026
-.031
.051
.067
.134
.00032
.002
.009
T-Ratio
57.546
6 854
-6.797
3 259
5.514
7 653
.712
.026
.476
.999
ermany
56.084
11.579
-11.64
.761
-.459
1.039
4.911
5 304
.227
.558
3 292
In addition to the variables listed, these models include controls for all the
variables found in Table 10.2, with the addition of interactions between (age-25) and
upper prof/mgr, lower prof/mgr, routine nonmanual, and self-employed.
t ion and training hav e imp orta nt , s ignificant direct effects on ou tcom es in
Germany despi te the t ight l inkage between educat ional and work ins t i tu
t ions .
To explore fur ther the re la t ionship betw een educa t ion and earnin gs
chan ge, w e inc lude d interac t ions betw een ag e and e duca t ion in the an alys is
of earnings
levels.
1
Table 10.6 sh ow s signif icant interac tions be tw ee n
educat ion and age in both countr ies , even when interac t ions between
occup at ion an d age are inc luded in the mo del . The interac t ions in bo th
countr ies show the same form: higher education has a bigger effect on
ea rni ng s for olde r tha n yo un ge r m en. If this effect is a life-course effect, the n
i t cer ta inly contradic ts the hypo these s of th is pap er , wh ich pre dic t th a t the
direct effect of ed uca tion in G erm an y shou ld shr ink quick ly as the s ta nd ard
ized creden tials in G erm an y are con verte d into occu pation al po sit ion s. An
al ternat ive in terpre ta t ion, how ever , i s tha t the in terac tions be tw ee n educ a-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 232/375
225
tion and age reveal cohort effects, rather than life cou rse effects. In this
interpretation,
the
effect of higher levels of education were greater for o lder
coho rts than they are for younger cohorts. Such an interpre tation is
consistent with a queuing interpretation of the value of creden tials (Thurow
1975;
Sakom oto and Pow ers 1994). In older cohorts, higher levels of
education are less common (e.g., Mare 1980, Blossfeld 1987), and so the
value of higher levels of education may be greater.
To im prove our ability to distinguish cohort from life-course effects, we
included interactions between age and education in our m odels for earn
ings chan ge. These models which are reported in table 10.7, reveal a
different story than do the models where earnings level is the dependent
variable. In the Am erican
case,
there is no significant effect of the interaction
between education and age on earnings change. In Germ any, the inclusion
of the education-age interaction reveal significant educational effects on
earnings change even in the second period, where the simp ler models found
no evidence of educationa l effects. In both G erman m ode ls, the interactions
imply a
decre sing
effect of educa tion on earn ings through the life
course.
It
is
tempting
to
interpret
this
declining effect as evidence that
job
position
is—
like the theory predicts— increasingly m asking the effects of educa tion over
the life course. However, this explanation ap pears to be inconsistent w ith
the da ta, since the declin ing effect of education on earnings change athigh er
ages
is
found even in reduced form regressions tha t omit all measures of
job.
In fact, we are reluctant to interp ret the greater effect of educa tion on recent
earn ings change as a life-course phenom enon . Instead, we suspect that—
as Blossfeld (1987) has argued— Germany s rigid labor market institutions
segment the labor market by cohort more completely than do American
labor m arke ts, and the pattern found in the recent data a re a reversal of the
cohort effects found in the model for 1985 earn ings. Additional research
with data that cover a longer time period is necessary to evaluate this
conjecture.
Tables 10.6 and 10.7 differ from table 10.2 in that they are regressions of
earnings change rathe r than of earnings at a poin t in time. Because the
dependent variable is change, we included controls for earnings level in
1985 on the change between 1985-88 and we included a control for the size
and direction of the 1985 88 change on the 1988 1991 change. Any tendency
for earnings change in one period to be offset by an opposing earnings
change in the second period constitutes an earnings reboun d. Table 10.8
shows that the earnings rebound was substantial in both countries, though
som ewhat larger in the United States. The possibility, indeed the fact of
earnings rebound in both countries raises the question whether the inci
dence and extent of such rebound is influenced by education . To exam ine
this question , we included an interaction betw een educationa l level and the
size of the earnings rebound in the regression for 1988-1991 earnings
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 233/375
216
TABLE 10.7 Effect of Age-Education Interac tions on Earnings Change
United States
1985-1988 1988-1991
Age
A ge
2
High School
Som e Col lege
College Degree
Extra Certification
High School (Age-18)
Som e College *
College Degree
(Age-25)
* (Age-25)
-.0088
.0001
-.037
.0248
.0578
-.0057
.003
-.0002
-.0004
-1.6
1.0
-0.9
0.7
1 5
-0.4
1.5
-0.1
-0.2
.0006
-.0001
-.0033
.0389
.1076
.004
.0033
.0015
-.0017
0.1
-1.1
-0.1
0.7
1.9
0.2
1.2
0.5
-0.5
Germany
Age
A ge
2
Lower-In termedia te
Upper - In t e rmed ia t e
Univers i ty
Lower-In ter (Age-18)
Upper - In t e r
(Age-18)
Univers i ty (Age-25)
Vocat ional Training
Fachschule
New certification
-.0041
.00004
.0139
.1443
.131
-.0004
-.0048
-.003
.0075
.0115
.0664
-1.3
1.2
0.6
3.2
4.0
-0.5
-2.3
-1.6
0.9
0.7
1.8
.00637
-.0001
.067
.0858
.1655
-.0024
-.0039
-.0082
.0051
.06
.0033
1.4
-1.5
1.8
1.3
3.7
-1.7
-1.4
-3.4
0.4
3.3
0.1
In add ition to the variables listed, these models include controls for all the
variables found in Table 10.2, with the addition of interactions between (age-25)
and upper prof/mgr, lower prof/mgr, routine nonmartual, and self-employed.
cha ng e. Specifically, w e included an interaction betw een the low est ed uca
t iona l categ ory and ed ucatio nal level. If w ork ers wi th hig he r levels of
education who suffered earnings reversals in the f irs t per iod were more
likely to experience earnings gains in the second per iod, then this interac
t ion term would be negat ive .
Table 10.8 sho w s that this term is instead insignif icant in both G erm an y
an d the United States . W hile hig he r edu catio n has a posit ive effect on
earnin gs and on earnings gro wth in both coun tr ies , it doe s not app ear to
pro vid e s ignificant ad van tage s for those worke rs w ho d o exper ience ea rn
ings decline. O ne possible exp lanation h as to d o with the reason s for
earnin gs decl ine . Earnings decline can occur through severa l mech anism s:
redu ction in ho urs w ork ed, redu ction in wa ge or salary in the sam e job, or
job chan ging. There is no par t icular reason w hy educa t ion wo uld be
associa ted with tha t par t of earnings rebo un d tha t come s f rom reduct ion s
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 234/375
217
TABLE 10.8 Ea rnin gs C ha ng e, 1985-1988, Inc lud ing
Intercept
Age
Age
2
H igh School
Som e Col lege
Col lege Degree
Extra Certification
Lower- In t e rmed ia t e
Upper - In t e rmed ia t e
Univers i ty
Vocat ional Tra in ing
Fachschule
New Tra in ing
U p p e r P r o f / M g r
Lower Pro f /Mgr
Rou t ine Nonman
Self-employed
Manua l Supe rv
Skil led manual
Empl . Tenu re
Un ion Member
Term Cont rac t
L o g H o u r s / m o n t h i n b a s e
Light Mfg
C o m m / T r a n / U t i l
W h o l / R e t T r a d e
Services
Log Earn88/Earn85)
Log Ear n88/E arn8 5) if +)
of Past Earnings Change
year
Log Ea rn88 /Ea rn85 ) Low Educat ion
Log Earnings in the 1985
R
2
Uni ted Sta tes
1.71
.004
-.0001
.073
.071
.096
.003
—
—
—-
—
•—
—
.107
.057
.039
.117
.095
-.017
.003
.027
—
-.182
-.036
.014
.008
-.002
-.549
.481
.105
-.075
0.13
6.2
0.5
-1.4
2.5
2.2
2.7
0.2
—
—
—
—
—
—
2.8
2.0
1.1
2.6
2.1
-0.6
2.0
1.2
—
-5.2
-1.3
0.4
0.3
-0.1
-10.7
6.4
1.0
-4.1
Effects
G e r m a n y
.829
.005
-.0001
—
—
—
—
.008
-.002
.021
.005
.049
.038
.054
-.007
.009
.012
.022
-.007
.001
-.007
-.047
.032
-.018
-.047
-.065
-.024
-.316
.161
.051
-.147
0.10
4.2
1.2
-1.8
—
—
—
—
0.6
-.1
1.0
0.5
2.7
1.0
2.6
-0.4
0.4
0.4
1.0
-0.5
0.9
-0.8
-1.5
1.0
-1.6
-2.9
-3.1
-1.8
-3.5
1.5
0.7
-7.7
in
hours or reductions in wages
in
the same
job.
One might
argue,
how ever,
that education should be most useful to reverse earnings declines that are
associated with job changes. To further exp lore the connection be tween
education and earnings rebound, we included an interaction between
education and earnings rebound for those respondents who changed
em ployers during the period of earnings declined. We still found n o
evidence that the rebound in earnings depe nds upon educational level.
12
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 235/375
2 8
Discuss ion
The results of this paper suggest that labor market s tructure shapes the
nature of the connect ion between educat ion and career outcomes , though
not a lwa ys in w ay s predic ted by theory. G erm any an d the Un i ted S ta tes
have quite different educational and training insti tutions, labor markets
tha t d if fer in im por ta nt w ays , and v ery dissimilar linkages betw een edu ca
t ional an d wo rk insti tutions. De spite these differences, indiv idua l- level
edu catio nal differences ha ve rather s imilar imp acts on earn ing s. Relatively
more of the effects of education appear to be mediated through occupa
tional posit ion in G erm an y th an in the United States w hich is in l ine w ith ou r
theoretical predictions. However , s ignif icant direct effects remained even
in the Germ an
case.
The m ed iatio n of edu catio nal effects in G erm an y by job
pos i t ion was most app aren t in the 1985 earnings regress ion. Such m edia
t ion was less vis ible in the earnings change models , par t icular ly when
interac t ions between age and edu cat ion are inc luded. O n the other h an d,
the sh ar p differences in the effects of vo catio nal train ing cred en tials, and the
large difference in the role of employer tenure, were evident through the
analyses presented here .
Cross-national contrasts are par t icular ly s trong during the 1988-1991
pe riod . The fact that edu catio n playe d a greate r role in the U.S. du rin g the
recession f i ts the interpretation of the American labor market as a place
where individual resources are par t icular ly impor tant whi le pos i t ional
resources are com parat ive ly unim po r tant . It shou ld be note d, how eve r ,
that the most inf luential educational difference in the United States con
cerne d the difference betw een ha vin g a hig h school dip lom a or not. The re
is no functional equivalent to this group in Germany, which is a pr incipal
reason w hy the G erm an ed uca tional effects tend to be sma ller than the
A m erica n ones. Similar ly, there is no functional e qu ivalen t in the U nite d
States to the training credentials of Germany, some of which continued to
have an impact on earnings change even in the 1988-1991 period.
Finally, the analyses in this pa pe r sh ow ed ed uca tion 's effects even in th e
indiv idua lis t ic Am erican labor m arke t to be less than pe rva siv e. W hile
education clear ly affected both earnings levels and earnings change in the
United States , education did not offer any special advantage to those
A m erica n wo rkers w ho se recent earn ing s trajectory w as in decline. W ork
ers w ho h ad lost gro un d in the recent pas t tend ed to m ake up a por t ion of
their losses . The po rtion that they m ad e u p (arou nd 55 in the U nited
States , 40 in G erm any ) w as the sam e on ave rage regard less of the level of
edu catio n. In other w ord s, edu catio n offered attractive possibil i t ies for
earnings growth regardless of the recent his tory of a worker 's earnings
trajectory. But edu catio n did not imp rov e a w ork er 's chan ces of m ak ing u p
lost gro un d. It m a y b e that m ore targeted training could serve this function,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 236/375
2 9
but such information was not available in the PSID, while the German
results offer little support for this position.
In su m m ary , the t ight ly s t ruc tured labor m arkets of G erm any differed
subs tantial ly f rom the m ore loosely s tructu red Am erican m ark ets in several
ways - in the importance of training credentials , in the unimportance of
em ploy er ten ure , an d m ost obvious ly in the overa ll sha pe of the e arnin gs
trajector ies , w hic h mo ve d in a relatively cohesive direction in G erm an y an d
div erg ed in the U nit ed States. A gain st this ba ck dr op , the effects of
indiv idua l- level resou rces ha ve m ore pro m ine nt effects in the United States
than in Ge rm any . H ow eve r , indiv idua l resources c lear ly m at ter even in
t ight ly s t ruc tured labor m arkets such as G erm any 's . These resul ts caut ion
agains t underes t imat ing the heterogenei ty of employment s i tua t ions tha t
exist in a complex economy, whatever the s tructure of i ts inst i tutions.
otes
This research ha s been suppo rted partly by National Science Founda tion G rant
S S 92-09159, partly by a g rant from Duke University's Arts and Science Council,
and partly by a grant from the Trent Foundation. This research was ma de possible
by a contract with the Deutsches Institut fur W irtschaftsforschung to use the
English-language ve rsion of the German Socio-Economic Panel. Please direct any
com men ts to the au thors at the Departmen t of Sociology, Duke University, Box
90088, Durha m, NC 27708-0088.
1.
The first w ave of the GSOEP was collected in 1984.
2. This pa pe r's use of recent panel data and its inclusion of foreign w orke rs w ho
reside in Germany (Auslander) distinguish the sample used in this paper from the
sam ples used in other analyses of earnings mobility in Germ any (e.g., Carroll and
Mayer 1986, Hannan, Schomann and Blossfeld 1990).
3. W e would havepreferred to use a measureof labor force experience rather th an
age,
but it is difficult to construct an experience estimate comparable to the
conventional age-education-6 me asure used for American data.
4. We used age rather than labor m arket experience in these results because the
differences in the education and training institutions of the United States and
Germany ma ke it difficult to construct comparable im puted measures of experience.
5. That portion of manufactur ing /m ining which fit Stinchcom be's (1979) defini
tion of large scale eng ineering (this includes mining) was grou ped in to the
omitted category.
6. The coefficients from this analysis are not included in the paper—they are
available from the authors upon request.
7.
The coefficients from this analysis are not included in the paper—they are
available from the autho rs upon request.
8. It is of course possible for nominal earnings to stay level or rise while real
earnings declines.
9. With this measu re, the regression coefficients measure the percent change in
the ratio of earnings at the two periods to a change of one un it in the covariate.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 237/375
22
10 . Relat ive to high school grad ua tes, college gra du ate s had
3 %
h igher earn ings
gr ow th in both the fi rst and second period in the United States. The contrast betw een
hi gh s choo l and college is significant at the .05 level in the first per iod , at the 0 1 level
in the second period.
11. To be precise, w e included interact ions be tw een h igh school com plet io n and
age-18, and betw een som e col lege or col lege deg ree com plet ion an d age-25. In
Germany, we inc luded in te rac t ions be tween lower school comple t ion or middle
school com ple t ion and age-18, and b e twe en u pp er school com ple t ion and age-25.
The cons tants are include d for conve nience of interpreta t ion. The resul t of this
approach is equivalent to the conventional method of incorporat ing labor force
exper ience (m easures of age - educa t ion - 6 ) un de r the ass um pt ion tha t ye ars of
school ing have no impact on earn ings over and abov e the four du m m y var iab les an d
the t raining variables inclu ded in our specification.
12.
These resul ts are avai lable from the authors upon request .
e f e r enc e s
Blossfeld, Ha ns-P eter. 1987. Entry into the Labor M arket and Oc cup at iona l Caree r
in the Federal Republic: A Co m par ison with Am erican Studies . Pp . 86-118 in
Com parative Studies o f Social Structure: Recent Research on
France
the United States
and the
ederal
Republic of Germany
edi ted by Wolfgang Teckenberg . Ar m onk ,
NY: M.E. Sharpe.
Blossfeld, Ha ns-Peter, Gianna Giannel l i and Karl Ulrich M ayer. 1993. Is Th ere a
N ew Service Proletariat? The Tert iary Sector an d Social Inequa l i ty in Ge rm an y.
P p . 109-135 in Ch anging C lasses: Stratification and Mobility in Post-Industrial
Societies
edi te d by Gosta Espin g-A nderse n. Beverly Hil ls , CA : Sage.
DiP rete , Th om as A. and Patricia A. M cM anu s. 1993. Te nure , M obil i ty , and
Incum bency: Co m parin g Observ ed Pat te rns of Earn ings w i th Pred ic t ions f rom
an Elaborated Theory of Occupat ional and Firm Labor Markets . Research in
Social Stratification and Mo bility
12:45-82.
DiP rete , Th om as A. and Patricia A. M cM anus . 1996. Inst i tut ions, Technical
C ha ng e, and D iverging Life Chan ces: Earn ings Mob il ity in the U.S. an d Ger
m any . Fo r thcoming in
American
Journal
of Sociology.
G a n z e b o o m , H a r r y B .G . , R u u d L u ij k x , a n d D o n a l d J . T r e i m a n . 1 9 8 9 .
In tergenera t iona l C lass Mobi l ity in Comp ara t ive P erspec t ive . Research in Social
Stratification a nd Mo bility
8:3-84.
H aller, M ax. 1987. Pos itional an d Sectoral Differences in Inco m e: Th e Fed eral
Repu blic , France, and th e Un ited States. Pp. 172-190 in
Com parative Studies of
Social Structure edi ted by W olfgang Tecken berg. A rm onk N .Y.: M.E. Sha rpe.
H ann an, Michael T., Klaus Scho m ann an d Hans -Peter Blossfeld . 1990. Sex an d
Sector Differences in the Dynamics of Wage Growth in the Federal Republic of
G e r m a n y .
American
Sociological
Review.
55:694-713.
Konig, W olfgang and Walter Mii l ler . 1986. Edu cat ional Systems an d Labor M ar
kets as Determinants of Workl ife Mobil i ty in France and West Germany: A
Co m par ison of M en 's Care er Mobil i ty , 1965-1970.
European
Sociological
Review.
2 (2):73-96.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 238/375
221
Mare, Robert D. 1980. Social Background and School Continua tion D ecisions.
Journal o f the American Statistical Association. 75:295-305.
Mayer, Karl Ulrich, David L. Featherman, Kevin
L.
Selbee, and Tom Colbjom son.
1989.
Class Mobility du ring the Working Life: A Comparison of Germ any and
Norway. Pp.
218 239
in
Cross National Research in
Sociology edited by Melvin L.
Kohn. New bury Park, CA: Sage.
Rosenbaum , James E., Takehiko Kariya, Rich Settersten and Tony Maier. 1990.
Market and Network Theories of the Transition from High School to Work:
Their Application to Industrialized Societies.
Annual Review of
Sociology.
16:263-299.
Sakom oto, Arth ur, and Daniel A. Powers. 1994. Education and the Dual Labor
Market for Japanese Men.
American Sociological Review
(forthcoming).
Thurow , Lester
C
1975. Generating
Inequality:
Mechanisms of Distribution in the U.S.
Economy. NY: Basic Books
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 239/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 240/375
Education and Credentialing Systems,
Labor M arket Structure and the Work
of Allied Health Occupations
Robert lthauser
Indiana University
Toby ppel
Yale University
The influence of professional sovereign ty on the division of labor in Am erican
medicine created fluid boundaries within the profession but . . . sharp
boundaries betw een physicians and o ther occupations. . . . Moreover, the
subord inate occupations, such s nursing and laboratory
work
became more
hierarchically stratified than did medicine. (Starr 1982:225)
Introduction
We have long known that s trat if ication within an organization or occu
pa tion can block mo bili ty, b y im pe din g access to jobs or occ upa tions with
m ore pay , respons ibi l i ty and auto no my . As S tarr sugges ts above, th is i s
readily ap pa ren t in the careers of those w ork ing in su bo rd in ate or al l ied
heal th (AH) occupat ions . Occupat ional s t ra t i f ica t ion within AH occupa
tions typically t ak es a du alistic form. For ex am ple , the re are tw o tiers in the
occupation of physical therapy: Physical Therapists (PTs) and Physical
Th era py As sis tants (PTAs). In this pa pe r , we will treat forms of s tratif ication
within a l l ied hea l th occupat ions— their educat ional an d credent ia l ing ins t i
tutio ns as well as their labor markets— as insti tution al obstacles to ind i
vid ua ls job mob i l i ty . W e wil l evaluate exist ing expla nat ions of b locked
mobi l i ty wi thin occupat ions and sugges t more cogent a l te rnat ives .
In the social s trat if ication l i terature, var ious proponents of functional ,
conflict and s tatus com peti t ion pe rspec tives gen erally try to exp lain restr ic-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 241/375
4
t ions on job mobil i ty by analyzing in different ways the consequences of
educational and other credential ing inst i tut ions (Bowles and Gintis , 1976;
Collins, 1971,1979). In the conflict and status competition perspectives in
part ic ular , the m ost relevant aspect of these inst i tut ion s is their intern al
s trat if icat ion, s tart ing with the fact that educational programs differ in
co nte nt , leng th an d d i f f iculty . Sim i lar s t ra t i f ica t ion is ap p a re n t in
credential ing or l icensing exams and their prerequis i tes . Applied to AH
fields , these perspectives view the relat ively high degree of occupational
stratification within AH occupations as a reflection of the degree of strati
f ication with in their edu cation al and crede ntial in g inst i tut ions. This sh ou ld
be especial ly app are nt in the relat ive diff iculty w hich those wo rk ing in the
lower t iers of each occupation have in accessing and completing educa
t iona l and credent ia l ing pro gram s.
In the du al labor m ark et l i terature, stratif icat ion with in labor ma rke ts is
the key source of blocked mobil i ty. Most of the dual labor market (DLM)
li terature focuses on the dualis t ic character of nat ional labor markets
(G ord on 1972; Berger and Piore 1980). Labo r m ark et sociologis ts and
eco nom ists ha ve largely neglected th e task of explain ing ho w labo r m arke t
dual ism develops wi th in spec i f ic occupat ions or organiza t ions (wi th im
po rtan t excep tions— Peatt ie 1974; Finlay 1983; O ster m an 1984, Parcel an d
Sickmeier 1988).
This pap er develo ps answ ers to these key ques t ions :
1. Is the blockage of mobil i ty in the work plac e th e inten de d, eve n
env is ione d effect of this edu catio nal an d creden tial ing s tratif icat ion?
2.
Is m obil i ty block ed for the reasons add uc ed by s tratif ication a nd labor
market theor ies?
Th e answ ers w e deve lop for these que st ion s wil l f inally lead us to a third
ques t ion:
3. What a l te rna t ive explana t ions have promise?
In our view , the internal s trat ificat ion of such al lied hea lth occ upa tions
as resp irato ry, physical and occu pation al ther ap y repres ent a t the ve ry least
a form of within-occupation labor market dualism. We firs t set out to
establish the plausibi l i ty of this view. Then, turning to the f irs t two ques
t ions abo ve, w e review the relevan t po rt ion s of functional and s tatu s
competi t ion theories of s trat if icat ion and of dual labor market theory,
specifying how they would explain this sort of labor market dualism. We
have drawn upon the expl ica t ions of func t iona l and s ta tus compet i t ion
theory pr ov id ed by Coll ins (1971,1979). Baron and H an n an (1994: 1121)
recently describ ed his 1979 vo lum e as the m ost influential w or k in sociol
ogy on educat iona l c redent ia ls in the labor market . . .
.
x
Th en, us ing the recent his tories of several A H o ccu patio ns as crucial data ,
we tackle the f irs t two quest ions above. As envis ioned in the creat ion and
im m edia te just if ica tion of educa t iona l and cred ent ia l ing pro gra m s, w e f ind
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 242/375
5
am ple evidence that a clear division of labor between tiers in the (typically)
hospital workplace was envisioned when the educational and credentialing
system s of each occupation were initially created. This norm ative expecta
tion of little or no overlap of responsibilities or duties between tiers of allied
health occupations was expected to parallel the larger, norm ative separa
tion of responsibilities between physicians and allied health workers. This
strict division of labor (or task d iscon tinu ity , to pa raphrase Offe (1976))
between tiers would result from and yet help reproduce the stratification
within the education and credentialing system and also provide the mecha
nism by which that stratification would effect blocked mobility in the
workplace.
However, our review of articles in the journals of these AH occupations
as well as our own field work strongly suggests that such an exp lanation of
blocked mobility and occupational dualism is inadequate. While upw ard
mobility from lower to upper tiers remains blocked in formal and some
times legal ways , we find permeable divisions of labor and far more task
con tinuity , e.g., overlapp ing responsibilities in the workplace, than wou ld
have been expected by the usua l stratification or du al labor market expla
nations. The stratification built into the educational and credentialing
institutions of these occupations does not engender and maintain a clear
division of labor, the theoretically expected mechanism for restricting
mobility in the medical workp lace. Pursu ing alternative explanations for
blocked mobility amidst this flexible division of
labor
we exam ine relevant
aspects of the histories of these occupations, isolating roles played by
occupational associations, hospital adm inistrations and by some key labor
processes, such as routinization and workplace assimilation .
Theoretical Background
To develop our argument more fully, we first need some background
from the literature on dual labor markets. In the initial writings on dual
labor m arke ts, w ha t Berger and Piore later term a minim alist (Berger and
Piore 1980:17 version of dualism was presented—a p rimary and secondary
marke t (Doeringer and Piore 1971; Gordon 1972; Edwards 1975). This
reflected the intellectual origins of DLM theory, with its initial focus on
urban unemployment and pover ty . I t sought to exp la in why
productivity-related factors like years of schooling or vocational training
failed to influence the employment of women or minority workers seem
ingly trapped in a sea of unstable, short-term jobs. The very name of this
concept suggests (recalling Stinchcom be s insight about concep ts as hy
potheses) its core hypothesis (Gordon 1972: 49-51): that there is little
mobility between dua l markets during individual careers.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 243/375
226
The Static Version of LM s
As concepts , DLM s are com mo nly u nd ers to od in very s ta tic te rm s. They
are por t ra yed as w ide divides be tw een pr ivi leged jobs— 'good ' , s table jobs
with careers—and underpr ivi leged jobs—'bad ' , uns table jobs wi thout ca
reers .
The core DLM hypothesis reinforces the image of an impermeable
bar r ier , unal tered b y chan ging econom ic c l imates or chan ges in techn ology
(such as adv anc es in scientific and m edical kn ow ledg e) . By the mi dd le of the
de ca de , Piore (1975:126) an d o the rs (Ed w ar ds 1979, G or do n et al. 1982)
found tha t d is t inc t ions wi thin the pr im ary mark et were as im po r tan t as the
dis t inc t ion between the pr imary and secondary labor markets , adding the
upper and lower t ier pr imary labor markets to the or iginal typology.
O n this s ta tic v iew, the 'bord er l ine ' be tw een m arkets is any thin g b ut
theoretically problematic. DLM theory thus presented fails to pay any
attention to the condit ions under which a l ine of work, a job, or an
occupat ion s foun d in one or ano ther market.
2
Un ders to od in th is w ay, dua l
labor market theory seems almost inapplicable to the present case of
internally s trat if ied occupations.
For tuna tely , wh a t underm ines th is com mo n s tat ic -mode l u nd er s tan d
ing are the arguments which appear (Doer inger and Piore 1971; E d w a r d s
1975) as par t of wh at R ose nbe rg (1989:365-366) calls a 'r icher v ers io n' of the
or iginal theory. He re , bo und ar ies w ere not v iewed as sta tic . Do er inger and
Piore (1971:169) do w np la ye d the 's tr ict sep ara tion ' of pr i m ary a nd secon d
ary labor ma rkets becau se they view ed the secon dary m arket as a m ix tur e
of internal labor markets ( ILMs) and non-ILM jobs. Craf t work was also
discu ssed as fal ling into its ow n labor mark et, a kind of ano m aly on the face
of the basic typ olo gy of labor m ar ke ts (Piore 1975:133-4).
ynamic Version of DLMs
A more dynamic unders tanding of labor markets over t ime becomes
apparent a t th is point , though much of the reasoning here has been ne
glected in subse qu en t em pir ical research. W hile no t specifically sensit ive to
internal s tra ti ficat ion within occupat ions , th is dyn am ic un de rs ta nd ing
provides us with the theoretical opening needed to establish that the
internal s trat if ication of AH occupations is an example of labor market
dual ism within occupat ions .
Th ere are two key s tar t ing p oin ts to this dy na m ic view . Th e firs t is that
labor m ark et se gm ents hav e different typ es of mo bili ty chains (Piore 1975:
129).
The second is that jobs can mov e back a nd for th across bo un da ry l ines
betw een the up pe r and lower pr imary or be tw een the lower pr ima ry a nd
se co nd ary tiers. D oerin ger an d Piore (1971) an d Piore (1975) disc uss at
length how jobs mig ra te be tw een m obi l i ty chains , hence ma rket segm ents
(1975: 139-148): . .. a par ticu lar job m ig ht lie on mo re th an on e type of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 244/375
7
mobility chain (Piore 1975:129). These autho rs also describe this as a
process of adjustment between em ployers' structuring of m arkets and
the supply of work ers w ith various behaviora l traits (Piore 1975: 147 ;
Doeringer and Piore 1971: 178f). From this reasoning w e can see how
physical therapy jobs, for exam ple, could be located in more than one labor
market segment.
The
third key aspect of
this
alternative view of
LMs
serves
to
correct the
atheoretical character of the usual literal or concrete interpretation of dual
markets. In his most extensive treatment of labor market dualism, Piore
asserts (Berger and Piore 1980:2,26-7) that the 'significance of dualism' is
not that markets are divided into two discontinuous segments, but that
markets are lumpy segments.
Whether or not there are two or more such lumps is not central to our
conception, though the numbers of segments cannot be multiplied indefi
nitely ...
This view is supp orted by the citation of findings from micro studies of
labor markets of 'dualisms within dualisms', e.g., Lisa Peattie's finding
(1974) of licensed and unlicensed street vendors in Bogota, previously
thought to be a homogeneous secondary market. This notion of more
microscopic d ualisms em bedded in more macroscopic dualism is consis
tent with ou r claims about the dualistic character of AH occupation s' labor
markets and of the more macroscopic labor market dualism dividing AH
and physician occupations.
Hence, a simplistic classification of AH occupations into one o r m ore of
these dualistic or triple segm ents
is
not our intention here. Indeed , inconsis
tencies mark the many attempts [at least seven, according to Rosenberg
(1989:369-371)] at placing occupations within the secondary m arke t or the
two tiers of the prim ary labor market.
3
Moreover, the occupants of the
therapeutic occupations und er study here are predom inantly wom en; yet
Piore adm its that this typology of segmented labor m arke ts was 'designed
largely to explain male careers' (1975:134).
Nor is our concern to establish the strictly dua l character of segm entation
within AH occupations. This would ignore the 'dualism within dualism'
insights mention ed above and w ould be doubtful in any event in light of the
occasional utilization of physical therapy aides, not to mention dental
assistants (versus den tal hygienists)
or
nurses aids working with LPNs and
RNs.
The point is simply that in the more dynamic understanding of DLM
theory appa rent in a careful read ing of the
literature,
market bound aries are
treated as problematic and subject to movement. Ongoing adjustments
con tinue at the boun dary. The literature even offers a theoretical opening,
rather than an impedim ent, to the treatment of within-occupa tional seg
mentation.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 245/375
228
The fact that most of the previous work on DLMs, especially but not
exclus ively by economis ts , has t rea ted occupat ions as homogeneous and
ent i re ly located w ithin the bo rde rs of some mark et segm ent is a con ven ient
bu t unn eces sary ov ersight. Most of the ear ly theoretical treatm en ts of du al
labor m ark ets failed to consid er the wa y in w hich occu pation s org aniz e a
great deal of work, or affect the character of whatever labor markets may
op erate in the wo rkp lace. The larger prob lem h ere is that for the m ost pa r t ,
theo ris ts ha ve lon g glossed over w ork plac e divis ion s of labor as a sou rce of
segmenta t ion among occupat ions , much less wi thin occupat ions .
There are four aspects of al l ied health (hereaf ter AH) work which
strongly support the presence of a dualis t ic form of labor market segmen
ta t ion
within specific AH occupations
The f irs t two are widely apparent to
ob serv ers of these occupatio ns an d will no t be fur ther arg ue d
h e re :
1 s t rong
w ag e differences be tw ee n th e tw o tiers of each o ccu pa tion (Visit2, p g . 15*);
and the blocked mobi l i ty be tween t ie rs .
5
The other two— inequal i t ies in
educational and credential prerequisi tes and the theoretical separation
betw een the tasks an d respons ibil i ties ass igned to each t ie r— are de velo ped
be low.
W e no w tu rn to the task of specifying h o w d ua l labor m arke t theo ry an d
functional and status competition schools of stratif ication theory can ex
plain this dualism.
Occupational Dualism:
The Account of Dual Labor Market Theory
For the purposes of ar t iculating a general account of dualism in ways
appl icable to occupat ional dual ism, we draw pr imar i ly upon the most
recent an d useful w ork o n du alis m the ory to da te (Berger an d Piore 1980).
At points , this mater ial extends the ear l ier writ ings in new ways; at other
points , i t restates arguments made ear l ier (Doeringer and Piore 1972,
Gordon 1972, Piore 1975).
Berge r an d P iore (1980:57-80) arg ue t ha t the labor m ark et loc ation of a job
is a function of job s tructure, which essential ly boils down to the number
an d com plex ity of tasks . Fu rthe r, the struc ture s of jobs reflect their lo cation
within a divis ion of labor . A continuum of divis ions of labor is imagined,
w ith the extremes of A d am Sm ith s pin factory and craft labor .
The sor t of divis ion of labor uti l ized, the argument continues, is a
function of the extent of the mark et (bo rrow ing from A d am Sm ith) , as w ell
as the s tabil i ty of demand and an abil i ty to divide product demand into
stable an d un stab le portion s (57-70). M uch of the evo lution of the job
s t ructure of ma nufactu r ing wo rk, par t icular ly lead ing to increas ingly spe
c ia lized jobs tha t occur red du r ing the r ise of ma ss prod uct io n, presu m ably
ref lects an expanding market. These jobs were pr imarily located in the
lower t ier of the pr imary market.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 246/375
229
In Piore 's ear l ier w ork , du al job s trata are pr incipa lly characte r ized b y a
contrast between 'general ' and 'specif ic ' behavioral trai ts required in their
work (1975; 130-134), and by differences in the types of mobility chains
found in each t ier
(128).
In his 1980 essa ys (17-18), Piore arg ue s furth er th at
...the list of distinc tive behav ioral characteristics w hich define the strata are
generated by single,
basic,
underly ing difference, that
is,
a difference in the
way in which people learn, and subsequently understand, the work which
they perform.
Ab stract lea rni ng typifies up pe r t ier jobs; con crete, on- the- job learn
ing is fou nd in low er- t ier jobs (20,70-78) . Abstract learning p erm its 'greate r
geographic and ins t i tu t ional mobi l i ty ' and is typica l ly acquired in educa
t ional sett ings. I t is then reinforced w he n ap plied at w ork in up pe r- t ier
jobs,
pr incipal ly through progress ions through upper - t ie r mobi l i ty chains ( re
turning on this point to his 1975 (133) treatment) . Concrete learning is
dev elop ed o n the job an d is prop or t ion ate to the nu m be r of tasks car r ied
out. W ha t is lear ne d is m or e specialized a nd less gen eralizab le to oth er jobs
or insti tutions. This difference in the 'way in which work is learned and
un de rs to od ' (1980: 21) establishes the differential l inka ge be tw een ed uca
t ional at tainment and labor market location.
W e ha ve now covered eno ugh of P iore ' s a rg um en t to isola te the essent ia l
pa r ts of a dualis t ic un de rst an di ng of the kin d of w ithin -occ up ation al s trata
w e see in A H oc cupa t ions and the ir wo rk. The role of a gro win g ma rke t in
st im ula ting a m ore specialized divis ion of labor w ou ld su gge st that
DL M hypo the s i s 1: a grow ing de m an d for services an d th e prosp ect of
a large s table dem an d for services w ou ld be a pr i m e condit ion for the
emergence of lower t ier AH strata .
W hile any s implis t ic association of 'abstra ct ' learnin g w ith the up pe r t ier
an d 'con crete learn ing ' with the lower t ier w ou ld face several obstacles , on e
would expect tha t
DLM hypothes is 2 : the higher level of educat ion and credent ia l ing
would include more theoretical or 'abstract ' t raining for the upper
than the lower tier .
Finally, the
DLM hyp oth es is 3 : wo rk of the upp er t ier occup ants shou ld requ ire
and uti l ize the more general skil ls and knowledge acquired in their
educ at ional pro gra m , whi le the wo rk of lower t ie r occu pants sho uld
uti l ize the more specialized knowledge acquired on the job.
In sho rt, DLM th eo ry offers three related gr o un d s for effecting the blo cka ge
of mobili ty between t iers . While we will not at tempt to directly test the
hy po the ses a bo ve, they do offer us a basis for projecting a ns w ers to our f irst
two ques t ions :
1. The blockag e of mo bili ty in the wo rkpla ce sh ou ld in de ed be th e
inte nd ed , eve n envis ione d effect of edu catio nal a nd cred ential in g s tratif ica-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 247/375
230
Hon with AH fields.
2. A blockage of mobili ty should be the result of several expected
differences between the tiers: the lower tier should utilize specialized,
on-the- job-based skil ls and knowledge rather than general and abstract
kn ow led ge , as par t of a bro ad er bu t fairly s tr ict divis ion of labor be tw een
t iers.
Occupational Dualism:
The Account of Two Stratification Theories
Functional Theory.
Fun ctional an d s tatu s com peti t ion theories differ
ent ly specify th is l inkage between educat ion and work, as many have
previously indicated (Coll ins
1971,
Berg
1970).
According to both accoun ts ,
the workplace is pr imarily a set of posit ions, of ten equated with dis t inct
occupations. But in the functional account, these posit ions require techni
cally de term ine d and objective skills . These posit ions or occu pation s differ
in level of skill , auth ority an d control , and in r ights and perq uis i tes (Davis
an d Mo ore 1945). Peo ple filling the m differ in na tiv e an d traine d c apa cities
and personali ty trai ts (Featherman 1975:333)
T h u s , functional theory sees a s tron g l inkage bet w een pro gre ss th ro ug h
educat ional ins t i tu t ions and both job placement and job a t ta inment . I t
a s sum es tha t
there re corresponding differences in the content of the specific work
done by those in these positions or occupations, e.g., a cle r technical division of
l bor
those
who
complete educ tion l
and other
credenti ls re pl ced
in positions
with d istinctly different responsibilities.
In functional theory, there is a s t ruc tura l i somo rphism betw een occup a
tional and organizational s trata and s trat if ied educational inst i tutions.
Access to work in different occupations and organizations is based on
comple t ing the r equ i r ements o f va r ious ly s t r a t i f i ed educa t iona l and
crede ntial ing insti tutions . Failure to com plete the requisi te level or d egr ee
of t ra ining and credent ia l ing expla ins subsequent immobil i ty be tween
different posit ions. Thus, despite i ts individualis t ic or ientation (Horan
1978), functional theo ry thu s assum es the existence of s trat ified edu catio nal
an d cre den tial ing in sti tution s, the focus of ou r firs t q uestion .
To sum up, b locked mobi l i ty is the unsurpr is ing consequence of the
un der lyin g educa t ional and credent ia l ing s tra ti ficat ion. A H pos i t ions exis t
w ithin a clear, technical divis io n of labor and ha ve their respective ed uc a
t ional and creden t ial ing prere qu isi tes . If ind ivid ua ls experience o bstacles to
mobili ty between lower and upper t iers , the ( theoretical) reason must be
that they have not progressed through all levels of the educational and
credential ing insti tutions within their AH f ield.
Collins S tatus Com petition Theory. H ow , according to Co l l ins ac
cou nt, do crede ntials typically lead to pos it ions w ith greater pr ivilege an d
greater rem un era tion , a feature even of the up pe r t iers of A H f ields? W hat
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 248/375
3
is the process, the mechanism ? Collins views pos itions as being a form of
prope rty, hence his concept of 'positional prop erty'. A 'pos ition' is a
collection of behavioral patterns . . . reserved for particular individuals
under particular conditions of tenure . (1979:53-4).
Ho w are credentials used to channel people? By conditioning access to
the best
jobs,
especially at their entry levels (1979:50-51). According to th is
theory, there are three m otives for restricting
access:
to control the training
process (those with the highest level of credentials gain more control over
their own reprodu ction); and to monopolize skill. Achieving these two
serves the third motive
—
to legitimate forms of privilege, greater pay , etc.
Exploring the first two of these motives, Collins views professional
monopolies as exam ples of Weberian 'sta tus grou ps '. Their basis is the
practice of certain esoteric and easily monopolized skills and the use of
procedu res that by their very na ture work most effectively throu gh secrecy
and idealization (1979:134). Earlier
(132),
he explains that a strong profes
sion 'requires a real technical skill that produces dem onstrab le results and
can be taught.
nly
thus can the technical] skill be
monopolized,
by controlling
who will be
trained. (Our italics)
But how are privileges legitimized in the credentialist view? As part of
his discussion of the historic evolution of 'education as status c ultu re' and
'education as
mechanism of occupational placem ent', Collins (1971:1015)
notes that: W ith the attainm ent of
mass . .. higher education in mod ern
America, the idea or image of technical skill becomes the 'legitimating
cu ltu re ' . . . [defining the] strugg le for position [that] goes on. In other
words , wha t legitimates inequality of privilege (in position and pay) is the
appearance of greater skill and knowledge possessed and exercised in
practice by those most highly credentialed. And this appearance is not just
in the eyes of the public or the law, but in the beliefs of those in the
workplace . This appearance is consistent only with a clear division of labor
between different positions.
To sum up , in Collins' status com petition theory, various work organi
zations and w ell-organized professions or occupations utilize credentials to
achieve more privileged and better paid positions by restricting access to
entry. They accomplish this by gaining and maintaining control over
training and hence their own social reproduction , monopolizing their skills
and thereby legitimating their level of privilege and pay . Their a ssum ptions
are embodied in pub lic and legal jurisdictional settlements; both assum e a
clear division of labor.
Collins' concept of 'positional prope rty ' helps define the charac ter of this
clear social division of labor. In his discussion of ways to dismantle the
'credential society', Collins (1979:198-202) argues that
the
key step wou ld be
to break the 'current forms of positional property', ending the clear divi
sions of labor that keep those in the occupations of m anager and secretary,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 249/375
3
or doctor or nurse, from sharing tasks and responsibil i t ies and uti l iz ing
on-the-job learning of skills to the fullest extent. Such a breakdown is no
m ore im aginable f rom a s ta tus com pet i t ion than a funct ional is t perspe ct ive .
But the flexible division of labor observed blurs the assumed line of
de m arc atio n in w ays an d to a deg ree wh ich is no t consis tent w ith the ten ets
of conflict theory.
A convergence of functional and status competition theories? Despi te
the i r d i fferences , we hav e sho w n tha t these theor ies share co m m on ass um p
tions about the central i ty of posi t ions and their location within a clear
division of labor, even as they offer differing accounts of how positions in
the work place are linked to educat ion . Acco rding to both v iew s , pos i t ions
differ in the skills they require. In status competition theory, these skill
requ irem ents of pos i t ions are m ore soc ia l ly cons t ruc ted and nego t ia ted; in
functionalis t theory, posi t ions require unequal amounts of technical ski l l
an d kn ow led ge . In the f irst , the clear divis ion of labor is social ly defin ed; in
the second, posi t ions are embedded in a technical divis ion of labor.
So for funct ional is ts , the com ple t ion of req uire d ed uc at io nal an d
creden t ia l ing ent ranc e requ irem ents br ing s access to pos i t ions . For Col lins
(1971),
s ta tus grou p mem ber ship cer ti fied by educat iona l and othe r formal
credentials lead to ini t ia l placement and later job mobil i ty. How posit ions
differ is variously described: in levels of skil l , s tatus honor, power and
rem un era tio n. Variat ions in these job characteris t ics reflect un de rly in g an d
clear differences in the con tent of work pe rfor m ed, w he the r this divis ion of
labor is social or technical in character. For both theories, positions are
clearly s tructu ral ly differentiated, as are dis t inct o ccup ations .
Applications to a History of AH Fields: An Overview So unders tood ,
the sha red logic of these theoret ical perspectiv es can be app lied to the labor
market dual ism apparent wi th in AH occupat ions . Such dual ism involves
tw o sets of jobs with very l imited mobil i ty bet w een each set. M ark et
du al ism m ay be based on the assu m pt ion of exper t , cul tura l auth or i ty found
in a professio nal m od el , or on the assu m ptio n of responsibil i t ies for con trol
an d di rec t ion found in the m ana gem ent
m o d e l .
In ei ther case, one finds du al
labor market s t ruc tures grounded on a d iv is ion of more f rom less exper t
labor.
The view tha t such a d iv is ion is app rop r ia te has long guide d the found
ing an d con t inuat ion of educat ional a nd cred ent ia l ing sys tems in genera l .
In the case of AH fields , this anticipated divis ion of labor w ithin A H w or k
is assu m ed in the very construc tion of a s trat ified ed ucatio nal system , w ith
i ts separa te educat ional t racks of unequal dura t ion, non-cumula t ive cur
ricula and l imited transfer of credit hours from lower to upper tracks. In
som e instanc es , there are also two d is t inct levels of vo lun tary certif icat ion
or s tate l icensing exam s. These typically pr esu pp os e an d often req uire prio r
com ple t ion of educat ional prerequ is i tes f rom accredi ted schools , prov iding
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 250/375
33
a relatively tight linkage between educationa l prepara tion and the right to
take the credentialing exam.
The application of this line of thinking
to
the boun dary between medicine
and allied health is already familiar. Physicians educational and medical
licensing systems share w ith allied health systems a comm on purpose— to
gain control of specific tasks and hence defend themselves against po ten
tially com peting occupational g roups, including (in the AH case) com peti
tion from lower tier workers . For various groups of phys icians, increasing
their own educational prerequ isites and later their voluntary credentialing
efforts (including the creation of the specialty certification) ha s enhanced
such controls and achieved insulation from competing medical and allied
hea lth occupations as well (Stevens 1972:198-200;
Starr
1982:356-7;
Abbott,
1988:
84). The basic dualism of physicians and allied health occupations
results.
Ironically, as allied health occupations em erged, the m odel prov ided in
this hierarchical differentiation between physicians and allied health occu
pations and their separate educational and licensing tracks was ad opted by
the leaders of allied hea lth occupations, as they developed similarly strati
fied educational and certification
tracks. s
part of this de liberate im itation
of these phys ician licensing and medical specialty board s, allied health
occupations have developed their own dualistic structure as a means of
formally identifying the elite members of their occupational group and
controlling the activities of subordinate m embers (Gritzer and Arluke
1985:
166).
The apparently more intense subordinate-level stratification Starr
noted (1982:225) cou ld be understood in part as a byproduct of an imperfect
and more rigid replica of the basic physician—allied hea lth division. This
is
reminiscen t of
DLM
theory s portrayal of education s greater im portance
for upper tier prim ary m arket mobility
chains,
and
its
correspond ing lesser
importance for low er tier primary market jobs (Piore, 1975:133).
Reviewing projected answers to key questions
This discussion of two
stratification theories suggests a second set of projected answers to our first
two questions:
1.
Is The blockage of mobility in the workplace the intended, even
envisioned effect of educational and credentialing stratification?
ANS: Though by different logics, according to both theories, the block
age of mobility in the w orkplace is the intended , even envisioned effect of
educational and credentialing stratification within AH fields.
2.
Is
mobility blocked for the reasons adduced by stratification and labor
market theories?
ANS: For both theories, an ind ividu al s potential for up ward mobility
is limited because they do not complete various levels of educational
prog ram s and c redentialing exercises that
are
prerequ isites for
the
upper or
lower tier positions of AH occupational positions; and because indiv idua ls
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 251/375
234
are pla ced ac cord ing to their cred entia ls in specific jobs w ith distin ctly
different responsibil i t ies , within a s tr ict divis ion of labor between upper
an d low er tie r w ork resemb ling similar d ivis ions betw een phy s ic ians an d
AH occupat ions .
Before turn in g to ou r data to check these projected an sw ers to the se first
tw o ques t ions , we firs t need to prov ide som e bac kg rou nd on the em ergenc e
of the second tiers of specific allied health occupations.
Data
O ur informat ion is dra w n f rom tw o types of sources . W e hav e review ed
relev ant ar t icles across several decad es of hea lth occup ation journ als on the
topics of education, cer t if ication, the application of new technology, and
ch an ge s in the s ize an d sc opes of practice. Beginn ing in 1986, the first a ut ho r
conducted informal interviews and observations of respiratory, physical or
occ up ation al therapists in m ore tha n 20 hos pitals located in seve n states ,
one Canadian province and two Australian s tates . In addit ion, the s taff or
elected lea ders hip w as intervie we d d ur in g vis i ts to the nationa l offices of all
three associations.
6
Extens ive in terviews we re con ducte d du r in g one
s ix-day 986 ann ual conv ent ion of the Am er ican Associa t ion of Re spira tory
Care and tw o annua l meet ings of the Indiana chapter .
Occup ational Dua lism in Allied Health Occupa tions:
Some Background
H ow d id occupation al du alism dev elop ? Let us consid er the low er t iers
of three al l ied health occupations: cer t if ied occupational therapy techni
c ians , physical therapy assis tants and cer t if ied respiratory therapy techni
c ians ,
and to a more limited extent, the case of licensed practical nurses.
Histor ically, the second t iers of physical , occupational and respiratory
the rap y formally dev elo pe d in the late Fift ies or du rin g the Sixties: Certif ied
O ccu patio nal Th erap y A ssis tants (COTAs) beg an in 1958 (Cro mw ell 1974),
Physical Th erap y A ssis tants (PTAs) in 1967, an d Certified Re spirato ry
Th era py Techn icians (CRTTs) in 1969 (only eight years after the ap pe ara nc e
of the upper tier Registered Respiratory Therapists or RRTs). In all three
fields,
on- the- job trained nonprofessional aides were used before and are
sti l l occasionally used now.
Certified Occupa tional Therapy Assistants CO TAs). In occup at ional
an d ph ysical thera py , the lower t ier or second level of the rap y w ork w as a
creatio n of their respe ctive occ up ation al association s. By the 1950s, the
en try level for occu pation al thera py (OT) w as a bac helo r s de gre e. Schools
were accredited by the American Medical Association, but credential ing
wa s un de r the control of the AO TA (OTs w ere not yet l icensed) . Th e AO TA
m a in t a ine d a r e g i s t r y ope n t o t hos e w ho ha d g r a du a t e d fr om a n
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 252/375
235
AM A-accredited p rogram and passed an examination. There was a strong
link between education and credentialing: graduation from an accredited
program was necessary to become a member of AOTA; also, a high
percentage of those who took the exam ination passed it. (Gritzer and
Arluke 1985)
A proposal was first made to the American Occupational Therapy
Association (AOTA) in 1949 to train OT assis tants in psych iatric hospita ls,
bu t it took nine years of debate before the COTA was created. Recognition
by the AOTA of
a
second level of worke rs in occupational therapy began in
a limited way in 1958. A shortage of personnel and outside pressure to
expand services we re the immediate stimuli. The AOTA at first app roved
a plan to certify and train w orkers in psychiatric settings only. Two years
later, AOTA perm itted parallel training program s for general practice only.
Finally, in 1963, program s trained assistants for all workplaces. These
programs were originally set up in the work sites for those already em
ployed as aides. In 1964, with the ap proval of AOTA, the training of COTAs
began to shift
to
junior
colleges.
AOTA both certified aides and accredited
educational program s
to
train
them.
C ertification was initially available for
those with work experience only; after educational standards were set,
training became a prerequisite. No examination w as required u ntil 1977.
COTAs were allowed to join AOTA, but for a long time they were not
eligible for office nor allowed
to
vote (Hiram a, 1986:23-24; Crom well, 1974)
OT assistants w ere originally to be used for ordering supp lies, prepa ring
equipment and maintenance, and 'diversion' (or therapy via crafts)
(Adamson and Anderson,
1966).
A 1968 Guide for the Supervision of the
Certified Occupational Therapy Assistan t stated that a COTA was to be
supervised by an OTR, an experienced COTA, or an OTR designate.
Physical Therapy Assistants
Physical therapy (PT), like occupational
therapy, had achieved a bachelor's degree entry-level to the occupation. As
in
OT,
PT schools were accredited by the AMA. Membership in American
Physical Therapy Association required graduation from an accredited
school. By the 1950s, physical therapists were in the process of breaking
down the domination of physicians through their support of new state
licensing
laws.
T herapists
in
some states were already engaging in indepen
dent practice. State licensing eventua lly led to the closing of the
physician-maintained Registry in 1971 (Gritzer and Arluke, 1985: 132).
The American Physical Therapy Association (APTA) resisted the move
ment to create a second level longer than did the AOTA. W ith the post-w ar
increase in rehabilitation
cases,
a shortage of PTs was ev ident. Non-registered
workers, or
T aides,
were used in most work places (Blood,
1965).
APTA's
official policy was to encourage on-the-job train ing of aides in specific work
settings, but to disapprove of separate courses for aides or assistants that
had physical therapy in the title or that would award a certificate.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 253/375
236
(W orth ing ham , 1965; Fow les and Yo ung , 1965).
By
1964,
w hen extend ed d ebate over es tabl ishing second t ier be gan , the
worker shortage had become cr i t ical . With Medicare on the horizon, the
de m an d for rehabi li ta tion w ork w ith chronic care pa t ients was expe cted to
dramat ica l ly increase . I t w as general ly agreed am ong PTs tha t no t e no ug h
new bachelors- level PTs could be recruited to satisfy the growing demand
(W orth ing ham 1965: 112-113). Federal reports were call ing for m or e
m an po w er in PT. Federa l laws ha d been enacted to pro vid e col leges and
trade schools with funds to set up training programs for al l ied health
w orkers . Junior and co m m un ity col leges , cons t i tu t ing a s t rong lobby , we re
eag er to obtain the federal funds to orga nize the training prog ram s. Th us,
there wa s con siderable external pre ssu re on PTs to extend services. Th e
lead ers felt that if they d id no t endo rse a lowe r level of wo rker , som e othe r
gro up w ou ld u nd er tak e to t ra in wo rkers outs ide of the jur isdic tion of PTs
(Wor thingham 1965:115) .
In 1967, after m uc h sp ir i ted deb ate, the APTA auth orize d a seco nd level
cred entia l (the PTA) to be bas ed on educ ation (White 1970). Pr og ram s w ere
im m edia tely set u p in junio r colleges. Un like OT , there was no gran dfa the r
clause enablin g those trained on the job to acqu ire the ne w crede ntial . PTs,
l ike RN s, w ork ed to al ter s tate l icensing law s to pro vid e for two levels w ith
sup erv ision of the low er by the up pe r . PTAs w ere m ad e aff il iate m em be rs
of AP TA to prev ent them f rom es tabl ishing a com pet ing org aniza t ion , bu t
given subordinate s ta tus and denied vot ing pr ivi leges .
Licensing laws for PTs were to specify that PTAs apply therapy proce
du res only un de r the di rec t ion and supervis io n of the regis tered PT
(Ru tan 1968:1001). The responsib ilities to be assigne d to PTA s w er e clearly
limited. As Fowles and Young (1965:124 - 6) advocated:
Using nonprofessionals in all of clerical, housekeeping, and m essenger dutie s,
limiting patient contact to the area of assisting the patient only when being
supervised or directed by the physical therapist...By adhering to [this and
othe r] simple principles,...the danger of less skilled workers carrying out the
physical the rapy techniques w hich are the PTs responsibility could be elimi
nated.
Of a l l the AH occupat ions , phys ica l therap y has ma na ge d to kee p its tw o
tiers furthest apart, at least on a formal level.
Licensed Practical Nurses
For the sake of comparison we offer this
rather br ief synopsis of the development of LPNs ( loosely drawn from
Kalisch a n d K alisch 1995:358-362; M elosh 1984:482-496). A s an early allied
health occupation, nursing was the ear l iest group to develop a dualis t ic
stru ctu re. RN s h ad bee n licensed since early in this cen tury . In the late 1940s,
most nursing education s t i l l took place in hospital diploma schools , but
there were some col legia te programs awarding a bachelors ' degree . The
leaders of the American Nurses Association (ANA) hoped that the entry
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 254/375
37
level of nurs ing prac t ice would eventual ly become a bachelors degree .
After the Secon d W orld W ar, ho w eve r , nu rsi ng faced a cri tical sho rtag e of
wo rkers . A l tho ug h some t ra ined prac t ica l nu rses pred ated W or ld War I I,
the great surg e of schoo ls for LPN s took place be tw een
1948
an d 1954. There
w as s t ron g outs id e pressu re , f rom surg eon s , for exam ple , for hosp i ta ls to
train auxiliaries.
Previously, the nursing leadership had resis ted trained auxil iar ies as
potential competi tors to RNs. But af ter the war they conceded to the
demand, endor sed the LPN, and sought to con t ro l s tandards and the
divis ion of labor be tw een RN s and LPN s. The Am er ican N urs es Associa
t ion (ANA ) lobb ied to rew rite s tate nu rsin g laws to inclu de LPN s, w ith the
provis ion tha t they be superv ised by RNs. LPN s we re cons idered n on pr o
fess ional workers and not a l lowed membership in the ANA.
Question 1: Wa s Blocked Mo bility the Intended Effect
of Educational Stratification?
Even before we reach the rather different case of Respiratory Therapy,
our review of the creation of a second t ier within nursing and especially
w ithin occup at ional and phys ica l therapy a l ready sugges ts an af fi rmat ive
answer to our f irs t question.
O T, PT and LPNs: A Comm on Scenario; A Deliberate Linkage.
W h e n
their dual systems of education and credential ing were f irs t created, nurs
ing, PT, an d OT ha d severa l fea tures in com mo n:
1 m em be rshi p in the occup at ional associa tion was l imited to those w i th
recognized educa t ion and /or c r eden t ia l ing ;
2) a severe shortage of workers was perceived which the occupation
could not f i l l based on i ts current s tandards;
3) up pe r t ie r m em be rs dom inat in g the associa tions s t rongly res isted the
notion that lower t ier occupants would receive any formal education or
training, other than in-service hospital training (Fowles and Young, 1965:
125-6;
Kirchm an an d H ow ard 1966, Gr i tzer, 1981: 275; W or th in gha m 1965;
Reverby 1987:163-6);
4) there we re low level tasks that seeming ly detrac ted from professiona l
status that might be delegated to lower- t ier workers , to the benefit of the
u p p e r tier (Perry 1964; Blood 1965);
5) although all f ields had on-the-job trained auxiliaries, the association
leaders wan ted a n ins t i tu t ional ized w ay to crea te mo re t ra ined w orke rs ,
fearing e i ther tha t o ther occupat ions doing lower- t ie r w ork w ou ld em erge
bey on d the reach of the occupa t ion s control or tha t adm inis t ra to rs w ou ld
substi tute lower- t ier workers (with their lower wage levels) for upper t ier
w or ke rs (Gritzer an d A rlu ke 1985:52, 71,77,116).
O ur rev iew of the l i terature in PT and OT jou rnals sh ow s us that the
founders of the COTA and PTA levels envisioned that a clear divis ion of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 255/375
238
labor would preva i l . Rout ine and supervised w ork w ou ld be conf ined to
oc cu pa nts of the new low er t iers ; other tasks wou ld be reserv ed for the
up pe r t ier . Som e wer e reluctant to delegate a ny tasks . In 1980, a directo r of
a PTA pro gr am rei terated th at the ' roles an d functions ' of PT an d PTA are
clearly sep arab le , an d that the role of the PTA will no t ch ang e in the next
de ca de . (Canan 1980)
The occu pation al associat ion, in each case, created or end or se d a s econ d
tier w hic h the up pe r t ier could con trol . The up pe r tier con sider ed i tself th e
prof ession al level; the low er t ier w as labeled the techn ical or ass is-
tan t level. (In OT especial ly, a con tinuin g batt le bet w een C OT As and OTR s
centered on the nam e tob e g iven the second
level).
Certification or licensing
was to remain under the control of the upper t ier .
Staff m em ber s a t the na t iona l hea dq uar te rs in both the Am erican Phy s i -
cal Th erapy Associa t ion and Am erican Occupat iona l The rapy Associa t ion
rei terated this view . Assis tan t OTs (COTA s) are no t seen as jun ior level
thera pis ts . . . This profession believes there is a dis t inct ion be tw ee n w ha t
the therapis ts do and assis tants [do] . (V isitl , pg . 9). . . . . there is a real
difference in the amount of responsibil i ty, independent function, cri t ical
thin kin g and pro ble m solving am on g PTs vs . PT As. (Visi t2, pg . 19)
Certified Respiratory Therapy Technicians CRT T): Different Scenario;
Same Result. Resp iratory ther apy (RT) w as a second gen erat io n all ied
hea lth f ie ld w hic h em erge d in the la te 1940s. By com pa riso n w ith the OT
and PT fields, the creation of the CRTT in respiratory therapy took place in
a som ew hat d i ffe rent context. A combinat ion of un usu al condi t ions prec ip i -
tated the creation of a technician level certification:
1) In the early 1960s, schools— which ha d just be gu n to be accredited by
the AM A— we re s t il l in hospi ta l se t t ings . A separa te ly incorp ora ted regis t ry
board, with physicians holding the majori ty of votes , was es tablished in
1960.
H enc e, the edu catio nal and , for the m ost par t , the cred entia l ing
sys tem s in RT w ere no t und er the control of the occ upa tional associat ion,
known ini t ia l ly as the AAIT (American Associat ion of Inhalat ion Thera-
pis ts ; and more recently as the AARC—American Associat ion for Respira-
tory Care).
2) Origin ally, the creden tial which the regis try conferred, the R egis tered
Res piratory ( init ia l ly, Inhalat ion ) Therap is t or RRT, w as exceedin gly diffi-
cult to obtain. Well over half of those w ho gra du ate d from accred ited
schools could not pass the writ ten and oral regis try examinations, though
the pass rates varied widely over time (Smith, 1989: 99,173,177.230)
3) M em bersh ip in the AAIT, unl ike the APT A an d the AOT A, was b ased
in the 1960s on w or k experience and n ot on gr ad ua tion from an accred ited
school.
4) The de m an d for respiratory thera py s ha rply increased in the la te 60s,
enhancing the perception of shortages of both RRTs and the more rapidly
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 256/375
39
growing segment of on-the- job trained technicians.
Th us the associa t ion w as num er ica l ly dom inate d by those wi tho ut any
crede ntial at al l , m ost of them traine d on the job (Smith, 1989; Bu rton an d
Ba rham , 1970). By 1969, w he n th e CRTT wa s created, there w ere a bo ut 7000
m em be rs of the occu patio nal association bu t only abou t 1600 RRTs (Smith
1989:
230).
6) Finally, the AAIT w as repe ated ly rebuffed in i ts effor ts to pers ua de the
registry, dominated by physicians and f i l led out by RRTs, to create a
cred ential in g m ec ha nis m for tech nicians. (Smith 1989: 95-97).
So the AA IT bo ldly created their ow n technician cer t ification m ech a
nis m , in pa r t as a poli t ical act aime d at the ph ysic ian -do m ina ted registry
board and in par t as a means of producing more workers and providing a
larger portion of exis t ing workers and members of the AAIT with a usable
crede ntial . For a short t ime (1969 - 1972) the AAIT adm ini ste red this
precursor to the current CRTT examination. The exam was open at f irs t to
m em be rs of the AA IT w ith w ork ex perience only (Smith 1989:80,116); i t wa s
relatively easy to pass. But the AAIT came to see that it had a conflict of
interes t be tween expanding i ts membership and giving exams to i t s mem
be rs .
In
1974,
faced w ith the com petit ion from AA ITs certificate, the re gistry
elected to ass um e the responsibil i ty for the technician level ex am a nd also
to reduce the physician dominance of the registry by reconsti tuting i ts
bo ard w ith equal phys ic ian and respira tory therapis t represen ta t ion (Smith ,
1989:79,97,139) .
The minimum length of therapis t educat ion was eventual ly se t a t two
years an d the leng th of trainin g for the technician at on e year . Th us t he re
is less of an ed uc ati on al difference be tw ee n levels than in OT or PT. (Smith
1989:
71-73; Burton and Barham, 1970).
Even in th is an om alo us case , the AAIT whi le crea t ing the CRTT w an ted
to pre serv e the hig her , Registered Resp iratory The rap y (RRT) cred entia l . I t
tr ied to establish dis t inctions between the tasks for which the two levels
w er e p re p ar ed (Ju lius, 1970). But of all the A H fields dis cu sse d he re,
respiratory therapy experienced the greatest diff iculty in establishing a
credential-based hierarchy in the workplace. .
Imm obility and Educational Stratification:
Sum marizing Answ ers to Question
#1
Let us sum u p the answ ers these m ater ials sugge st to ou r f irst qu estio n,
is the blockage of mobi l i ty the in tended ef fec t of educat ional and
cred entia l ing s trat ification? Oc cup ation al leaders expe cted that s tratif ica
t ion within educat ional and credent ia l ing ins t i tu t ions would shape work
place s trat ification an d block mo bili ty be tw een t iers . This w ou ld be accom
plished by restr ict ing access to upper t ier credentials in two ways. Firs t ,
mobi l i ty from lower to up pe r tie rs w ou ld be ha m pe red becau se access to
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 257/375
24
the exams that certify qualifications for jobs at either level was (and
remains) limited to those who have completed accredited educational
prog ram s. For example, in respiratory therapy,
the
right to take credentialing
exams has typically depe nded on completing educational prerequisites at
accredited programs and in some instances, months to years of practical
experience (Smith 1989: 97,168,178-9).
Second, academic credits gained in completing the lesser levels of
training eventually established for lower tier degrees hav e rarely counted
against the requirem ents for upper tier degrees. So experienced lower tier
occupants wh o
go
back to school' while wo rking part or full time find they
m ust largely 'start over', a situation which is often lamented at all levels of
these occupations (Kerr 1971).
As
a result respiratory therapy's two-track educa tional tiers, each tightly
linked w ith two-track credentialing
exams,
com bined to make career mobil
ity difficult, even 'virtually nonexistent' in the early period of 1969-1973
(Smith 1989: 97, 165) and still difficult now . Indeed, PT educators have
criticized high school counselors for misleading graduates into thinking
that the associate degree program for
PTAs is
a stepp ing stone for entry into
the professional pro gram . (White 1972).
In effect, those guid ing the creation of assistant level prog ram s with their
very separate educational or certifying tracks viewed
the
work tasks central
to each of the two levels as distinctly different, or 'discontinuous'. Their
underlying m odel of working practice and app ropriate training a nd educa
tion envisioned very little continuity across the whole range of tasks
unde rtaken by upper and
a
lower tier
AH w orkers.
In their vision, functions
sharply differ between the levels of this simple hierarchy—a prim e exam ple
of what Claus Offe has defined as a task-discon tinuous statu s organiza
tion . (Offe 1976:25-27) This model also rem inds us of the sep ara tion of
con trol from execution, of thinking from doing,
a
key princip le of 'scientific
management ' .
In short,
the restricted mo ility
in this occupational context (which recalls
that of a classic dual labor market) is the intended consequence of a
considerable degree of educational and credentialing stratification, itself
pred icated on a vision of a clear division of labor and a task discontinuous
organiza tion of the statuses and ap propriate w ork for each tier. What this
explanation leads us to expect in practice is a clear division of labor, with
highly skilled tasks reserved for upper and lesser skilled for lower tier
workers. But is this the actual workplace arrangement which produces
blocked m obility? Is the way in which mobility is actually blocked consis
tent with this explanation based on educational and credentialing stratifi
cation systems?
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 258/375
24
Question #2 : Is Mob ility Blocked for the Reasons Adduced
by Stratification and Labor Ma rket Theories
If al l ied hea lth f ields w ere s tructu red to imitate, ho w ev er ap pro xim ately ,
the m edica l profess ion 's emph as is on credent ia ls and edu cat ion, one w ou ld
expect a conformity or even an over-conformity to the practice norms
accompanying this (greater) internal s trat if ication. The work typically
assigned to the upper t ier should uniquely require their higher levels of
educat ion and cer t i f ica t ion; converse ly , less demanding or respons ible
w ork w ou ld be cons igned to lower t ie r incum ben ts . W hat is the po int of
having levels of medical occupations subordinated by cer t if ication and
edu catio nal cr i ter ia if no t to s truc ture the str ict sep aratio n of m ore rou tine ,
from less, discretion-filled from discretion-less tasks, on the basis of ac
cepted qualif ications?
But observ at ions d o not conf irm this expecta t ion. While the c reden t ia l ing
sys tem m ay be an ins t i tu t ional extens ion of the educat ional sys tem s, ne i ther
is at all consistent with the flexible division of labor found in actual
occ up ation al practice. H ow this com es ab ou t is relevant to ou r appra isal of
the arguments drawn from dual labor market and s trat if ication theories .
In actual practic e, acco rding to a great deal of our ma ter ial f rom inter
v i e w s ,
observation and ar t icles in the occupation's journals , the scopes of
practice of workers in the two t iers substantial ly overlap. We have found
that in many though not al l hospitals , lower t ier occupants of the var ious
du al labor m arke ts enco mp ass ing several sub ordin ate medical occu pat ions
a re
not
uniformly exclusively assigned to less skil l-demanding, discretion-
exercis ing tasks. For exam ple, in respirato ry c are units in ho spitals , m ost of
the hospital ICUs we observed or heard about could not s taff ICUs around
the clock with RRTs while confining CR TTs to general f loor du ties . In som e
hospi ta l depar tments , PTAs and COTAs develop highly ski l led areas of
specialized care, bec om ing the care-givers of choice as well as co nsu ltants
to their upper t ier colleagues.
These observations are consis tent with mater ial we found in these
occupations ' journals . Task overlap among different levels of AH occupa
t ions has been a m ply d oc um en ted (Golds te in an d H orow itz : 1974: 21-38;
Cro mw ell 1968; A da m so n an d A nd erso n 1966). The over lap of skil ls an d
respons ibi l it ies m us t be recog nized in PTs w ork (Holmes 1970); W at ts
outl in es factors w hic h m ak e f lexibil ity in deleg ation of responsib il i ty
hig hly de siro us (1971:32-3). The skil l levels of PTA s m ay far exc eed thos e
for PTs in the pe rfor m an ce of certain task s. . . (W hite, 1970; see also W atts
1971: 33). O ne s tu dy d oc um en ts tha t over one th i rd of the w ork ing t im e of
the ear l ier PT aides wa s spen t do ing lo we r skil led tasks (Blood 1965:120) .
Another shows tha t a ides were working outs ide of the areas they were
or iginal ly expected to wo rk in and mu ch of the t ime w ere wo rking w itho ut
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 259/375
4
the expected sup ervis ion by PTs (Gray
1964).
Anothe r au tho r warn s aga ins t
PT l icensing law s tha t specify edu catio nal subject m atter o r dutie s and task
a l loca tions , because 'de ve lo pm ent in care techniq ues ' may m ake an y g iven
divis ion of labor and educational content obsolete (Rutan 1968). Over the
his tor ies of these fields it
is
app aren t tha t A H f ie lds in genera l and the u pp er
t ier workers in part icular have s imilarly ass imilated many aspects of
trea tm en t w hich w as formerly the exclusive pro vinc e of ph ysic ian s . (Perry,
1964:434; Melosh 1984:492).
W ha t this and oth er evi den ce d ep icts, in short, is close to w ha t Off e (1976)
ca lls " task- con t inuo us" s ta tus organiza t ion . As exempl i f ied by the opera
t ions of a small craft workshop, functions, ski l ls and knowledge overlap
with in the f ramew ork of a progress ive dev elop m ent of sk il ls , know ledg e
and responsibil i t ies
7
This is illust rated in the follow ing excer pt from an
inter vie w wit h a phy sical therap y assis tant (PTA):
Ques.: W hat are the differences in the work done by a PTA vs. a FT? What
actually hap pen s? Ans: When
an
assistant goes to school, they co ncentrate
on
how mo dality is done, how to administer treatm ents. Therapists concentrate
more on evaluating the problem and some therapists also carry out the
treatm ent. The basic difference is: we do n't e%
r
a uate how to take care of a
patient.
We can
change treatmen ts after discussion with the therap ist.
We
can
change or upd ate the program if
necessary.
Qu es: Do you have to do that—
discu ss it w ith the therapist—first? Ans: Kinda..., sorta. We use our own
judgement: say if a patient is not getting relief then we can try another
modality. This is an assessment, based on our professional experience; we
reassess what the patient needs. These have to be informal; at the initial
evaluation, the
T
puts
a
patient through the tests—I do n't know the name
of
these tests, but these are part of the formal, initial evaluation. Assistants do
informal reassessments instead.
We
can do muscle grades, test their strength,
reassessments, similar to what was part of the initial tests.
While this passage shows expected differences in schooling and knowl
edge be tween PTs and PTAs, the PTA ques t ioned combines a res ta tement
of the no rm ativ e ass ign m ent of ini tia l assessm ents to PTs w ith the inform al
practice of PTAs doing reassessments .
These comments i l lus trate the overlapping scopes of practice of the two
t ie rs in the workplace , an i s sue over which respi ra tory and occupat iona l
therapy have exper ienced cons iderable turmoi l . A major ro le de l inea t ion
s tu dy of respi ra tory care w ork in the mid-Seven t ies found tha t a t 'beg innin g
prac t ice , [ there w as only] one leve l of pr ac t i t io ne r . . . [wh ether thera pis t or
technician, an indiv idu al perform ed]. . . the sam e set of tasks , i rrespectiv e of
edu cat ion a l back grou nd. . . " (Smith 1989:163). The resul t was an im po rtan t
poli t icalco ll is ionbe twee n the m em bers hip associat ion and the creden tial ing
bo dy , resul ting in a ch an ge (in 1983) in the con tent of the ex am s for the CRT T
an d RRT crede ntials an d the designa tion of the CRTT as the only entry level
exam.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 260/375
43
In the f ie ld of OT, the same issue has drawn much at tention. Before a
decade o f expe r ience wi th COTAs has pas sed , Adamson and Anderson
(1966, pg 78) re po rte d that:
An assistant wh o
is
hired to take the
place
of an
OTR is
frequently on the same
level in the organizational structure of the department as a staff therapist.
Even the text of a Guide for the Supervision of Certified Occupational
Therapy Assistants (1968:99) stated: Inheren t in the [superv ision] relation
ship is the acceptance... of the fact that readiness for increased responsibility
should when possible, be recognized by assignment to more demanding
duties and rew arded by an approp riate adjustment in salary scale.
As an im m edia te pas t pres iden t of the AO TA , Ruth B runyate fe rvent ly
restates sent im ents exp ressed at a recent (1968) w or ks ho p on the 'ass is tan t
movement ' be low; ye t even these ideas re f lec t a common ambiva lence
about clear versus flexible divisions of labor. This passage of hers comes in
an art ic le auth ore d and la rge ly wri t ten by the then current AO TA pre s ide nt
Florence C rom w ell (Cro m we ll 1968: 379-380):
subm it that the goal of m any voices here is to have as the pro fession's official
stand, the fact tha t the COTA
is
an occupational therapist and not
a
subservi
ent assistant to an occupational therapist; that with in the profession the re are
varying avenues of educational preparation, qualifying criteria and levels of
competence. I infer that we recognize that just as competence, capacity or
judg em ent and exercise of responsibility vary w ith the level of preparatio n, so
too do they exist with individual variation within either professional or
technical level...there can and do exist COTAs wh ose com petence for
a
given
assigned responsibility exceeds that of a more elaborately educated but less
endo we d professional, just as there do exist COTAs and OTRs wh ose indi
vidual competence falls below our desired optimum for an assigned task.
[OTRs mu st reassess their roles with assistants aroun d] . . . w ith the added
manpower provided by the COTA the profession is now experiencing a
reevaluation of wh at it does and d oes not, should an d sho uld not
do.
We now
acknowledge that contained within m any occupational therapy services are
great pockets of activity program s u nder the guise of occupational therapy
which d id not shock or even disquiet the profession, u ntil it saw that wh en the
COTA functions with an appropriate pattern of supervision, he is highly
skilled in perform ing pa rts of all facets and every stratum of the profession's
practice.
Somewhat more res t ra ined is the response of Florence Cromwel l , the
current AOTA pres ident , as she comple tes th is a r t ic le . The unse t t l ing
adjus tments which upper t ie r workers s t ruggle to make in the wake of
obvious ly o ve r lap pin g scopes of prac tice a re ap pa ren t as she w ri tes :
We of the OTR grou p, having helped to create this needed level of function
within o ur practice, also struggle with his role as his proficiency and potential
increase,
as
dem ands on us change,
as
we attempt
to
identify o ur own rightful
and most efficient function in today's practice.
In short, flexible d ivisio ns of labo r are evid en t in practice. This is far from
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 261/375
244
the vis ion of a clear divis ion of labor upon which the educational and
credential ing systems were grounded. In the face of blocked mobili ty
be tw een low er and up pe r t iers of these A H occu pation s and the inabil i ty of
the educat ional an d c redent ia l ing ins t i tu t ions to mater ia l ly control w ork
place practice, what explanations of blocked mobili ty make sense, which
also take into account these f lexible divis ions of tasks among upper and
lower tiers?
Question 3: Wha t Alternative Explanations M ake Sense?
Fu ture research in this area sho uld focus on alternative exp lana tion s. W e
d o not vie w a ny of the se as likely to be establis hed as the sole, be st acc ou nt.
Allowing for the inf luence of unique features of each occupation and i ts
w ork , a satisfactory exp lanatio n m ay w ell requ ire som e com bin ation of al l
of these.
Actions of Occup ational Associations.
O ccu patio nal associations, l ike
the AN A , AO TA , APTA or AARC, as wel l as respira tory th erap y 's ind ep en
de nt regis try and accreditat ion bo dies , clearly ha ve an interest in h o w labor
is divided between lower and the controll ing upper t iers . By dint of
m em be rsh ip rules and pa t terns of h is tor ic dom inance , grad uat es of u pp er
tier schools , posse ssors of the top level creden tials and o ccu pan ts of u p pe r
tier posit ions have consis tently dominated these associations (with the
possib le exception of som e per io ds of the recent his tory of the re spira tory
the rap y establishm ent) . These associations can take actions w hic h m ain tain
m arke t bo un dar i es and block mobi l i ty be tw een t ie rs. A m on g the type s of
actions that have been apparent, we br ief ly discuss three examples.
1.
Creating then S abo tagin g an Alterna tive Career Ladder for CO T
As .
Occupat ional Therapy is an AH occupat ion which has probably devoted
m ore effort and reso urces to creating non - t radi t ional aven ue into i ts up pe r
t ier than an y other . W ith some ongo ing amb ivalence , the Amer ican Occu
pat ion al Therapy Associa t ion und er too k a career mobi l i ty pr og ram in the
Sevent ies . CO TAs we re permit ted to take the exam lead ing to the regis t ry
on th e bas is of five or m or e years of exp erien ce, six m o nt hs of 'pro fessio nal
f ield work' and a 'self -assessment ' , rather than having to fulf i l l the usual
edu catio nal req uire m ent s (Visit l , pg . 1). The fact that suc h an en d- ru n
could be mounted reflects the realities of a workplace more flexible in its
divis ion of labor than formal dualism would be inclined to acknowledge.
W hy di d this 'mob il i ty p ro gr am ' fail? A relevant but secon dary factor in
its demise was that the response to this program was l ight; fewer than 50
CO TAs succeeded in becom ing OTRs (Ada ms 1981). Dou bts m ou nted wi th
the ensuing costs in t ime and money. More essential were doubts that the
OTR registrat ion e xam alone wa s sufficient to cert ify s om eo ne as an OTR,
abse nt the usu al edu catio nal pre requ isi tes . (Visi t l , pg . 8) But the fatal issue
w as no t the technical quali ty of the exam . Faced w ith even m od est
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 262/375
45
prospects of success , the most important reason for i ts demise was f inal ly
ar t icula ted: no am ou nt of exper ience ho w ev er se lf -s tudied , superv ised and
documented by th is program could by i t se l f replace formal educa t iona l
req uire m en ts . In effect, the exp erim ent had clarified and s tren gth en ed the
fundam enta l a s sum pt ions of educa t iona l dua l i sm .
OTRs at the s tate level a lso exercised their veto through l icensing laws:
the i r s ta tes dec l ined to recognize COTAs who became OTRs through the
program, s ince they lacked the cus tomary accredi ted educat iona l c reden
t ia ls for beco m ing an OTR. (Hig htow er V a n d a m m
1981;
Ry an 1 986:27-30).
The effect ive result was that a fai led nine-year experiment had served to
blun t the cha l lenge to OTR dom inat ion f rom w ha t some referred to as the
ass is tant m ov em en t (Brunyate in Crom wel l 1968). CO TA s hav e s ince
presse d for o ther forms of enh anc ed s ta tus wi th in the AOTA , bu t no longer
by the m ea ns of a job ladd er l inking COTA s and O TRs. The tens ion pers is ts
be tween the formal and informal pa t te rns t raced out above .
2 Raising minimum educational standards for upper tier recruits
While ca l ls for increas ing minimum educat iona l requirements for nurses ,
OT s or PT s can be fou nd in the his tories of these fie lds, the projected im pa ct
on workplace prac t ice of ac tua l ly achieving these minimums is ne i ther
obv ious nor as in ten ded . Dep end ing on the f ie ld , the in tent was to e l imina te
or at least diminish the use of lower t ier workers , or to enhance the
occu pat ion ' s s tandin g in the m edica l com m uni ty . Some leaders of respi ra
tory therapy, however , have argued tha t the e ffec t of a l ready higher
edu cat iona l s tand ard s for up pe r t ie r w ork ers in phys ica l and occupat iona l
ther apy h as been to enha nce the grow th an d use of lower t ier wo rkers and
therefore have opposed proposa ls to move respi ra tory therapy toward a
four-year mi nim um t ra in ing per iod . H ow eve r d i ffe rent ly the leadersh ip of
a given occupation 's upper t ier foresees the consequences of rais ing educa
t ional s ta nd ar ds , the div is ion of labor betw een t iers is l ikely to be affected
(Abbott 1988:84).
3 Delineating career ladders within the upper tier During the las t
deca de , the leade rship of Aus t ra l ian nu rs ing p ut th rou gh a res t ruc tur ing of
their conventional career ladder, which originally led from entry level
regis tered nursing to various supervisory posit ions (Silver 1986b). An
addi t iona l new t rack now permits nurses to remain heavi ly involved in
clinical care , even at higher levels of expert ise . Largely missing from the
documenta t ion of th is development was the impact which th is res t ruc tur
ing ha d on the sk il ls tha t the Au s t ra l ian ' enrol led n ur ses ' (ENs are ro ugh ly
equ ivale nt to LPN s in the U.S.) w ere al low ed to de ve lop a nd use on the job.
While this redelineation of career l ines was l imited to RNs and did not
di rec t ly reach down to ENs, the la t te r group exper ienced new pressure to
confine themselves to the more routine, discret ion-less duties of nurs ing,
especial ly in the ur ba n h osp ital set t ings . This is an in terest in g case of labor
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 263/375
246
market boundar ies be ing reconf i rmed by a renewed re inforcement of a
no rm ativ e, clear divis io n of labor , e .g., c lam pin g do w n on well es tablish ed
tendencies for second tier nurses to assume the tasks and exercise the skills
forma lly rese rve d to RNs (Jenkins 1989; Silver 1986a).
Hospital Administration.
As in this latter insta nce , the oth er gro up of
stakeholders with vested interests in how labor is divided are the hospitals.
Given the oppor tuni ty for observat ion, no one can miss the remarkable
degree of var iat ion in the ways in which hospitals organize the work of
all ied health o ccup ations. The s tr ict divis ion of labor found in som e de pa rt
m en ts of hos pitals resem bles the vis ion of task -disc on tinu ity foun d e ar l ier ;
in other departments , one sees the f lexible divis ion of labor even though
mixed with heavily restr icted chances for actual mobili ty. However na
t ional organiza t ions ap proac h jurisdic tional i ssues , hosp i ta l and dep ar t
m en tal adm inistra tors (and in som e cases, me dical directors) routin ely
create the kind of divis ion of labor they w an t in their units . Som e do this by
creating clinical ladders for personnel. While generally not designed to
byp ass educ at ional a nd credent ia l ing ins t i tu t ions or to chal lenge the block
age of mobili ty s t i l l engendered by educational and credential ing insti tu
t ions , these ladders do provide local s tatus and f inancial recognit ion of
growing ski l l and knowledge.
In shor t , occupat ional associa t ions and in dividu al hospi ta l adm inis t ra
t ions act to directly shape the scopes of practice among upper and lower
tiers of al l ied hea lth occu pation s, quite ap ar t f rom their no m ina l su pp or t of
established insti tutions of education or credential ing.
Workplace Labor Processes: Routinization and Assimilation . Abbott
(1988:59-65;84) arg ue s tha t con trol ove r task s is the cen tral o rg an izi ng
reality of profe ssion al life . Ind ee d, the issue of con trol ove r w or k is cen tral
to the sociology of work. The inabil i ty of credential ing or educational
insti tution s of AH occu pation s to control the al location of tasks in the hea lth
care wo rkplace rem inds us of the broa d arg um ent f rom s tud ents of the labor
proc ess l i terature: that social aspects of the labor process su ch as w or ke rs
strug gle (Claw son 1980), w ork er s craft kn ow led ge (Kusterer 1978) or
making out (Burawoy 1979) and other workplace games have genera l ly
im pe d ed a ttem pts of external actors to control or inf luence v ar io us asp ects
of wo rk— wh ethe r ma na gem ent or labor unio ns or here , educ at ional an d
credential ing insti tutions, occupational associations and hospital inst i tu
t ions .
Given the var iable l inkage betw een control over tasks and the
possibil i ty of job mo bili ty eviden t abov e, w ha t sor t of labor proce sses m ay
be helpful in explaining this strange combination of a flexible division of
labor coupled with blocked mobili ty?
Abbott argues that control of tasks or jur isdictions is pursued in three
arenas : the legal sys tem, publ ic opinion and the work place itself Sett lement
of these jurisd iction al claim s is subject to perio dic re neg otia tion , b ut this is
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 264/375
47
more frequently observed in the workplace than in the other arenas. In
explaining why, his wo rk identifies several related labor processes relevan t
to this alternative explanation.
As skill and knowledge develops and work practices evolve, a large
prop ortion of treatm ents, tests, assessments, and diagnoses simply becom e
routine. Abbott (1988: 126-7) who has already underscored the conse
quences of dividing professional work into its routine and nonroutine
elements, describes as 'deg rad ation' this shifting of increasing rou tine tasks
to other professional or occupational groups. At the level of everyday
practice, Abbo tt's concept of workplace assimilation specifically con
cretely cap tures the process of transferring know ledge an d skill and recalls
the spotlight on the im por tant role of 'craft know ledg e' (Kusterer 1978) in
the labor process literature. Halpern
1990)
has similarly described how the
'routinization of work' helps explain an intra professional jurisdictional
adjustment between pediatricians and social workers, psychiatrists and
pediatric nurse practitioners.
One interesting and relevant feature of the history of medical practice is
the successive transfer of procedu res initially reserved to med ical special
ists or ordinarily ph ysicians into the daily responsibilities of various levels
of allied health occupations. The passage of blood p ressu re m easure m ents,
venous blood sam pling (Melosh 1984:492) and m ore recently, arterial blood
gas
sam pling represen ts some examples of
this.
W hat is mo re pertinen t here
is that the same downward distribution of skills and knowledge, super
vised tasks and responsibilities occurs from up per to lower tier AH occupa
tions. Hence this passage eng end ers par t of
the
flexible d ivision of labor we
found above.
When groups of physicians delegate (or lose control over) routine
proce dures to allied health occupations, they presumab ly turn their atten
tion to new , non -rou tine activities, enhance their academic research inter
ests or otherwise hold exclusive jurisdiction over other tasks. Failing this,
they risk extinction. There is am ple evidence that the upp er tiers of AH
occupations envision moving themselves into new responsibilities when
they similarly 'hand off now routine tasks to lower tier members. The
bou ndar ies rem ain intact, howev er, whether between tiers of AH occupa
tions or betw een physicians and AH occupations. In this way, the ong oing
processes of routinization, workplace assimilation and degradation pro
vide the steady social undercurrent driving the transfers of now routine
tasks , consistent with the flexible division of labor we found above. In labor
marke t terminology, bou ndaries between
the
tasks norm atively assigned to
physicians, upp er and a lower tier AH w orkers are successively redraw n,
providing for the continuation of blocked mobility.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 265/375
48
Summary
We have argued that the upper and lower t iers of three al l ied health
occu pat ions represen t an exam ple of occup at ional dual labor m ark ets . We
have reviewed the way in which the ass is tant-level , lower-t iers of al l ied
heal th occup at ions developed as a m ean s of de term ining w he the r the i r
es tabl ishment and opera t ion have been cons is tent wi th dual labor market
functional and status competition theories of stratification.
De spite som e differences in their accou nts , these theories clearly antici
pa te a l inkage be twe en the educa t ional an d credent ia l ing sys tem s of a l l ied
heal th occupat ions , the i r p lacement of graduates of these sys tems in to
dis t inct ive posi t ion s in the wo rkp lace and their sub seq uen t job m obil i ty o r
im m obil i ty. Even after their ini tia l plac em ent in entry-level po si t ions, the
mechanism facil i ta t ing the blockage of mobil i ty between t iers is a clear
divis ion of labor be tw een phy s ic ians an d bo th t ie rs of a l lied hea l th occupa
t ion. En vis ion ing such a divis ion of labor, ne w ly trained ph ysicia ns , and
up pe r and lower t ie r A H incum bents hav e been prepa red for the ir respec
t ive tasks and responsibil i t ies by their experiences in their dis t inct ive
educat ional and credent ia l ing t racks .
Throughout much of the discussion found in the l i teratures of these
therap eut ic occu pat ions , how ever , and f rom m ost of our observa t ion s an d
interv iew s in the work plac e, the blockag e of mo bil i ty is no t acco m plishe d
by a consistent and strict division of labor. We finally considered several
factors as possible explan ation s of bo th the substan tial blockage of mo bil i ty
be tw ee n t iers and th e absence of s trict divis ion of labor a m on g t iers . These
included act ions by these occupation's t rade associat ions, their related
acc red i t ing and c reden t ia l ing bod ies , and workp lace p roces ses l ike
rout in iza t ion and downward ass imi la t ion of upper t ie r sk i l l s and knowl
edge .
i scuss ion
W e ant ic ipa te three counter a rgu m ent s to th is pa pe r and offer respo nses
to each. Firs t , some might argue that dualism within an occupation (or an
indu s t ry) so far mis repre sents the m ore agg regate form of dual ism across
occupations and industries as to comprise a tr ivial , not a s trategic, tes t of
extant theory . W e m ight hav e symp athized wi th tha t v iew before s ta r t ing
this l ine of research. The trouble is , the occupa tional fou nde rs of the low er
tier often write or talk as if they were schooled by a dual labor market or
s tratif icat ion theoris t W e w ou ld argu e that the divis ion s w ithin the
workplace which they envis ioned in the detai ls of their educational and
credent ia l ing ins t i tu t ions look very m uch l ike the bro ade r forms of du al ism .
That the flexible division of labor we and they find instead is inconsistent
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 266/375
249
with both extant theory and their own expectations is not a tr ivial fai l ing.
Secondly, o thers may view the com binat ion of task over lap an d bar r iers
to mo bili ty as entirely consis ten t with a credential is t acco unt w hich s tresses
the po w er inequal i t ies manifes ted in the cont inu ed m ainte nan ce of mob i l i ty
ba rr iers , ev en in view of the kind of pos it ional differentiat ion that al l three
theore tica l perspect ives assum e and ( w e think) require . Ho w ever , those
hav ing this v iew w il l hav e to specify other mecha nism s by whic h those w i th
credent ia ls hold sw ay than they ha ve previous ly specif ied. They cann ot
long evad e the pro ble m of how the legi timacy of credent ia ls is ma inta in ed
in workplaces wi th task over lap, whi le publ ic unders tanding and legal
codes seem oblivious to workplace reali t ies . We doubt a credential is t
account of unequal pr ivileges in the workplace can explain the fair degree
of task ove r lap be tw een the unequal ly pr ivi leged pos i t ions .
We have suggested that in the view of s tatus competi t ion theory, the
cred ent ialed eli te seek to legit ima te their pr ivileg e and to restr ict access into
their ranks by c la iming a connect ion betw een the ir super ior credent ia ls and
their skil ls, kn ow led ge a nd abil ity. Highe r- level crede ntials are an cho red in
licensing law an d in pu blic belief that the crede ntialed do ind eed h av e m ore
skil l or knowledge, as evident in the case of people who successfully
m asq ue rad e for years as doctors . Prop on ents of th is theory m igh t argu e tha t
their theory does not require these skil ls to be real nor must the claim be
factual ly cor rec t. Their un de rs tan din g of cont inu ed pr ivi lege and ap pa ren t
po w er migh t eve n enco mp ass a vers ion of false consciousn ess ( see espe
cially Bo urd ieu s ac cou nt in 1977a:496) . The indefinite lag wh ich A bbo tt
projected between init ial changes in workplace jur isdictions (here repre
sen ted by task overlap ) and later pub lic and legal resett lem ents of jur isdic
t ions leave t ime for a var iety of outcomes to develop, as suggested by
A bb ott s typolo gy of jur isdictiona l sett lem ents (Abb ott 1988:69-79).
C an sta tus g ro up pr ivileges be sustaine d on the basis of false belief in the
face of an unclear divis ion of labor be tw een gro up s? From the vanta ge po int
of sign alin g theo ry, this is a case of Ta reto -inf erio r or inefficient sig na lin g
equi l ibr ia base d on informat ional con s idera t ion s (draw ing here on Baron
an d H an na n 1994:1122). H ow s table is th is ar range me nt? Acc ording to
Baron an d H an na n, Spence (1974) views this type of equ il ibr iu m as f ragile ,
especial ly in the face of com pet i t ion an d exp er ime nta t ion wh ich ma y lead
to recogni t ion of a mo re sup er ior equi l ibr ium . We bel ieve a thoro ug h
trea tment of the his tory of occupat ions and profess ions would turn up a
var iety of ou tcom es, inc lud ing bu t not restr icted to a) a decline in the po w er
and numbers of upper t ie r workers (compare phys ia t ry in re la t ion to
ph ysic al th era pis ts, in the mid-1930s and late 1950s in H alp er n 1992), b) a
combinat ion of upper t ie r s de legat ing rout ine work to lower t ie rs whi le
moving themselves in to new areas of undisputed exper t ise and exclus ive
respons ibi l it ies (note ped ia t r ic ians po s tw ar shif t in to psychosocia l pe dia t -
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 267/375
250
rics, re lat ing actual care to non-physician groups in Halpern 1990:33-37) ,
c) a contractual sett lement designed to reestablish intra-occupational or
intra-profession boundaries in the face of high levels of task overlap. The
la tte r a re appa ren t in F inlay s (1983) account of wha t hap pe ne d w he n a
group of s teady but noncredent ia led crane opera tors were shar ing the i r
high payin g work wi th s teadily wo rking but credent ia led op era tor s . A ne w
labor contract supp or te d by both union an d emp loyer associa t ions placed
the noncredent ia led but essent ia l ly equal ly ski l led opera tors behind new
barriers to mobil i ty.
In our view, status competi t ion theory cannot be restated to effectively
ignore the legi t imacy problem or to predict the inevitable victory of the
ini t ial ly m ore pow erful , becaus e more highly creden tialed, t ier of wo rk ers .
A s ta tus com pet i t ion theory depe nd en t on publ ic ignoranc e w ou ld hav e to
predict that once the public decides that such beliefs are false, credentials
will be deleg i t imize d and the privi le ged status of the top tiers will end. Th e
w id es pr ea d aboli t ion of l icensure laws in the U nite d States in the 1800-1860
per iod demonstra ted jus t such a resul t : author i ty by l icense requi red an
asso ciation of licensing and objective skill tha t the pu blic an d th e law co uld
no t sustain (Starr 1982:58). Un less this theory can otherw ise solve the
problem of credentials legi t imacy, overlapping scopes of a practice pose
continued risks to public belief in the objective skill and credentialed
superiori ty of the upper t iers .
otes
1.
We do not draw upon the work of either Bourdieu or signaling theory (Spence
1974),
tw o theories which could be linked to a status com petition perspective. While
Pierre Bourdieu is also considered a status competition theorist, we will not deal
with this though t in this paper. His concept of hab itus or predispositions in many
respects parallels Piore s argum ent see below) about general
vs.
specific behavior
traits and the corresponding difference in the way people in the different tiers of
a
dual labor market learn from and understand their own work. However, we have
not found a treatment of workplace stratification (1977a,1977b) along dim ensions of
status (as opposed to social class) in Bourd ieu s work.
According to Rosenbaum s (1986) treatment of signaling theory, the general
context of signaling theory is the difficulty emp loyers have processing my riads of
job applicants or candidates for promotion, as they attempt to infer ability from
multiple pieces of information. While educational creden tials would app ear to be
a
central criterion for hiring or promotion, many other criteria are found imp ortan t,
according to Rosenbaum s review: job dem ands, as proxied by job status and
earnings (149), differences in family social background and track placement in
school
(150),
and the status of entry jobs (153) are others which come into play. The
career pattern predicted by signaling theory, in Rosenbaum s view, is irregular:
thus, he found no clear career structure at the utility company he studied (152), no
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 268/375
252
single selection criterion or no set of criteria could be linked to students' track
placement, etc. In these many respects, signaling theory would appear wholly
inapplicable to the AH case, without being fairly tested by the case
2.
Many of the often cited explications of dual labor market theory by Edwards
(1975,1979) and Gordon et al. (1982) reinforce this static understanding.
3. The empirical research (Osterman
1975
which first assigned occupations to the
upp er and lower tiers of the primary as well as to the secondary ma rkets placed all
of the incum bents of a given occupation into a single tier. For exam ple, 'the rapis ts
and he alers' was placed in the lower tier of the primary market. Setting aside the fact
that his method of placement
was
entirely subjective,
the
criteria he used for placing
occupations
into
markets were 'the degree of autonom y and personal pa rticipation'
(for assignments to e ither tier of the primary market) and low wages and skill levels
and unstable employm ent for assignment
to the
secondary
market
More systematic
is
the use of measures of repetitiveness and Specific Vocational Preparation from the
Dictionary of Occupational Titles by Spenner et al. (1982), which result in placing
'thera pists' and 'therapist assistants' in the upper tier of the primary market.
4. All citations to V isitl refer to interview notes taken during a day-long visit
on August 10,1990 to the AOTA offices near W ashington
D.C.;
those to Visit2 to
notes taken d urin g a visit on August 15,1990 to the offices of the APT A, also near
W ashington D.C.
5. To be sure, some aspects of AH work would also support locating AH jobs
within ILM structures: the on
- the
- job acquisition of add itional
skill
and knowledge
in hospital settings, not to mention the persistent shortages of labor (Althauser and
Kalleberg 1981; A lthauser
1989).
Yet very few ILM structure s for AH occupations
have emerged, despite some experiments to that end (Goldstein and Horowitz
1977).
6. These include the following groups: The American Occupational Therapy
Association, The Am erican Physical Therapy Association, The Am erican Associa
tion of Respiratory Care, the Joint Review Committee for Respiratory Therapy
Education and the Na tional Board of Respiratory Care, Inc.
7.
Offe (1976) further describes 'task-continuous status orga nizatio n' as involv
ing A greater mastery of
the
[common, technical] rules and greater ab ility, know l
edg e and experience in production.... wh ich differentiates those in the upper from
lower strata. He con tinues: ...there is a wid e area of technical rules to wh ich equal
obed ience is requ ired from all the occupants... of the adjacent positions.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 269/375
252
e f e r enc e s
Abbot t , Andrew. 1988 . The System of Professions. Chicago: Universi ty of Ch icago
Press.
A da m s, N.G. 1981. La dde r to Professional Cert i fication: The Car eer Mob il ity
P r o g r a m .
American
Journal
of
Occupational
Therapy
35(5): 328-331.
A d a m s o n , M a r g a re t J and M ary Alyce A nde rso n. 1966. A Study of the Uti l izat ion
of Occup at ional Therapy A ssis tan ts and Aides ,
American
Journal
of Occupational
Therapy 20 : 75-79.
Al thau ser , Rober t and Arn e Kal leberg .
1981.
Fi rms, Occ upat ions an d the St ruc ture
of Labor M arkets : A Concep tua l A nalys is .
P g .
119 -149 in Sociolog ical Perspectives
on
Labor
M arkets
edi ted by I . Berg. N ew York: Ac adem ic P ress.
A lthau ser, Robert . 1989. Interna l Labor M arke ts . Annual Review of Sociology
15:143-161.
Baron, Jame s N. an d Michael T. H an na n. 1994. The Impact of Econo mics o n
Contemporary Socio logy . Journal o f Econom ic Literature 32:111-1146.
Berg, Ivar. 1970.
Education and Jobs: The
Great
Training Robbery.
Ne w Y ork: Praeger .
Berger, Suzanne and Michael Piore. 1980. Dua lism and Discontinuity in Indu strial
Societies. Cambridge:Cambridge Univers i ty Press .
Blood, He len. 1965. Their Use by the Profession . Journal of
the
American Physical
Therapy Association
45:118-124.
Bou rdieu, Pierre . 1977a. Cu ltural Re prod uct ion and Social Re pro duc t ion . Pp
71-112 in Richard Brow n, ed. ,
Knowledge
Education
and C ultural Change.
Lo ndon :
Tavistock 1973. Rep rinted in Jerom e Karabel and A. H. Ha lsey. 1977. Po w er and
Ideology in Educat ion. New York: Oxford, pp. 487-511.
. 1977b.
Ou tline of a Th eory of Practice.
Trans: Richard N ice . Ca m brid ge and
New York: Cambridge Univers i ty Press .
Bowles, Sam and Herbert Gint is . 1976. Schooling in Capitalist America. New York:
Basic B ooks.
Burawoy, Michael . 1979.
Manufacturing Consent.
Chicago: Universi ty of Chicago
Press.
Burton, Ge orge C. [MD] and V irginia Z. Barham [RN, EdD] 1970. Ed uca t ion a n d /
or Licensure in Inhalat ion Therapy, Inhalation T herapy 15:143-149.
Ca nan , Betty . 1980. W hat Chan ges are Predicted for the Physical Th erap ist Assis
tant in the 1980s?
Physical Therapy
60: 312.
Clawson, Dan. 1980.
ureaucracy
and the
Labor
Process. New York: Mo nth ly Review
Press .
Coll ins, Randal l . 1971. Fu nction al an d Conflict Th eorie s of Ed uca tiona l Stratifica
t ion . American Sociological Review 36:1002-1018.
. 1979.
T he Creden tial Society.
N ew York: Acad emic Press .
Cr om w ell , Florence S. 1968. Na tional ly Speaking: from the Pre siden t . American
Journal of Occupational Therapy
22:
377-379.
. 1974. The Deve lopm ent of the Oc cupat ion al Therap y Assistant : History
and Sta tus Report .
Journal of the American Speech and Hearing Association.
16:
671-676.
Da vis, Kins ley an d Wilbert Mo ore. 1945. Som e Princ iples of Stratification . Ameri-
can
Sociological
Revieiv.
10: 242-249.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 270/375
253
Doeringer, Peter and Michael Piore. 1971. Internal Labor Markets and Manpower
Analysis. Lexington, MA: DC Heath.
Edw ards, Richard C 1975. The Social Relations of Production in the Firm and Labor
M arket Structure.
Pg.
3-26 in Richard
C
Edwards, Michael Reich and David M.
Gordon (eds), Labor Market Segmentation.
. 1979.
Contested Terrain.
New York: Basic Books.
Featherman, David L., Frank L. Jones and R obert M. Hauser. 1975. Assumptions of
Social Mobility Research in the U.S.: The Case of Occupational Status.
Social
Science Research
4:329-360.
Finlay, William. 1983. One Occupation, Two Labor Markets: The Case of Long-shore
Crane O perators.
American Sociological Review
48: 306-315.
Fowles, Beth H. and Janet E. Young. 1965. On-the-job T raining.
Journal of the
American Physical Therapy Association
45:124-126.
Goldstein, Harold M. and Morris
A.
Horow itz.
1977. Entry Level Health Occupations.
Development and
Future. Baltimore: Johns Hopkins U niversity Press.
Gordon, D avid. 1972.
Theories of Poverty and Underemployment.
Lexington, MA: DC
Heath.
Gordon, David M., Richard Edwards and Michael Reich. 1982.
Segmented
Work
Divided
W orkers. Cam bridge: Cambridge U niversity Press.
Gossett, Robert. W hat's the Answ er? Queries and Replies of Concern to Thera
pists. Physical Therapy
52: 95-96.
Gray, Jim Mason. 1964. Function of Nonprofesssional Physical Therapy P erson
nel. Journal o f the American Physical Therapy Association
44:103 -109.
Gritzer, Glenn. 1981. Occupational Specialization In Medicine: Knowledge and
Market Explanations.
Research
in
the Sociology
of
Health Care
2 251-283.
Gritzer, Glenn and A rnold Arluke. 1985.
The Making ofReh abilitation.
1985. Berkeley:
Univ. of California Press.
Guide for the Supervision of Certified O ccupational Therapy Assistan ts. 1968.
American Journal
of
Occupational Therapy
22: 99.
Halpem , Sydney A.
1990.
Medicalization as Professional Process: Postwar Trends
in Pediatrics .
Journal o f Health and Social Behavior.
31:28-42.
. 1992. Dy nam ics of Professional C ontrol: Internal Coa litions and
Cross-Professional Boundaries.
American Journal o f Sociology
94 (4): 994-1021.
Hightow er-Vandam m, M. 1981. Soaring into the 80s New Directions.
American
Journal
of
Occupational Therapy
35: 767-774.
Hiram a, Haru. 1986. The COTA: A Chronological Review in
The Occupational
Therapy
Assistant
Roles and
Responsibilities: A
Comprehensive
Text for
Practice and
Education
edited by Sally E. Ryan. Thorofare, NY: Slack.
Holmes, Thelma M. 1970. Supportive Personnel and Supervisory R elationships.
Physical Therapy
50:1165 -1171.
Horan, Patrick. 1978. Is Status Attainm ent Research Atheoretical?
American
Sociological Review. 43: 534-540.
Jenkins, Enid. 1989 Nurses' Control over Nursing . Pgs. 193-208 in Issues in
Australian Nursing 2
edited by Genevieve Gray and Rosalie Pratt. New York:
Churchill Livingstone.
Julius,
Louise H. 197 Depth of Knowledgeof Technician and Therap ist,
Inhalation
Therapy
15: 71-75.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 271/375
54
Kalisch, Philip A. and Beatrice J. Kalisch. 1995.
The Advance of Am erican Nursing.
Philadelphia: Lippincot t)
Kerr, El izabeth E. 1971. Uti l izat ion and Pre para t ion of Pers onnel to Deliver H eal th
Ca r e ,
Respiratory
Care 16:41-48.
Kirc hm an, Ma rgaret M and Billie H ow ard . 1966. The Role of the Cert i fied As sistant
in a Genera l H ospi ta l .
American journal of Occupational Therapy
10:29 3 - 297.
Kusterer, Ken C. 1978.
Know-Ho w on the job: The ImportantWorking Know ledge of
nskilled Workers.
Boulder, CO : W estview Press.
Me losh, Barbara. 1984. M ore than the 'Phy sician 's Ha nd ': Skil l an d Au tho ri ty in
Tw entieth C entu ry N urs ing . Pg. 482-496 in
Women and Health
Care
in America
edi ted by Judi th Walzer Leavi t t . Madison, WI: Universi ty of Wisconsin Press.
Offe, Claus. 1976.
Industry and Inequality.
London: Edward Arnold .
O ste rm an , Pau l . 1975. An Em pirical Stud y of Labor M arket Seg m entat ion .
Indus
trial and
Labor
R elations Review 28 :
508-523.
. 1984. W hi te- Collar Internal Labor Mark ets. Pg. 163 -190 in
Internal
Labor
Markets
edi ted by Paul Osterm an. Ca mb ridge M A: MIT Press .
Parcel , Toby and M. B. Sickmeier. 1988. O ne Firm, Tw o Labor Ma rkets: The Case
of McDonald 's in the Fast Food Industry. Sociological Quarterly. 29: 29-46.
Peat t ie , Lisa. 1974. The Other Sector: A Few Facts from Bogata, Som e C om m en ts
and a List of Issues . M imeo .
Perry, Jacquel ine. 1964. Professional ism in Physical Th erap y.
journal of
the
Ameri
can Physical T herapy Association
44:4 29 - 434.
Pior e, M ichael. 1975, N ote s for a Th eor y of Labo r M arke t Stratification . Pg.
125-150 in Labor
Market Segmen tation
edi ted by R. E dw ard s, M. Reich and Da vid
M. G ordo n . Lexington , MA : DC H eath .
Reve rby, Susan. 1987.
Ordered to
Care The
Dilemma of American Nursing 1850 -1930.
Cambridge: Cambridge Univers i ty Press .
Ro senb aum , James E. 1986. Inst i tut ional Car eer Structure s and the Social C on stru c
tion of Ability . Pp. 139 -171 in John G. Robinson (ed),
Hand book of Theory and
Research for tlie
Sociology
of Education.
Rose nberg, Sam. 1989. Fro m Segm entat ion to Flexibil ity . Labour and Society.
14:363-407.
Ruta n, Fred M. 1968. Imp licat ions of Ass istant-Training Pro gra m s for Physical
Therapy: On Licensure . Physical Therapy 48 : 999-1002.
Ry an, Sally E. 1986.
The Occupational Therapy Assistant Roles and Respo nsibilities: A
Com prehensive Tex t for
Practice
and Education.
Thorofare, NY: Slack.
Si lver, M argaret . 1986a. A Prog ram for Car eer Struc ture. A Vision Becomes a
Real i ty . Australian Nurses journal 16:44-47.
. 1986b. A Pro gra m for Career Structure . From N eo ph yte to Expert .
Australian Nurses
Journal
16:38-41.
Smith, Gary A. 1989. Respiratory
Care.
Evolution of a Profession. Lenexa, KS: Applied
Measurement Profess ionals .
Spence, A. Michael. 1974.
Ma rket Signaling: Informational Transfer in Hiring and
Related Screening Procedures. Cambridge . Harvard Univers i ty Press .
Spen ner, K enneth, L uther B. Otto, and V aug hn R.A. Cal l . 1982.
Career Lines and
Careers.
Lexington , MA: D.C. Hea th and Co m pany .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 272/375
255
Starr, Paul. 1982.
The Social Transformation
of
merican Medicine
New York: Basic
Books.
Stevens, Rosemary.
1972. merican Medicine and the Public
Interest New H aven: Yale
University Press.
Watts,
Nancy T. 1971. Task Analysis and Division of Responsibility in Physical
Therapy. Physical Therapy
5 1:
23-35.
W hite, Barbara C
1970.
Physical Therapy Assistants: Implications for the Fu ture .
Physical Therapy
50: 674-679.
. 1972. W hat's the Answer? Queries and Replies of Concern to Therapists.
Physical Therapy
52:
94-95.
W orthingham, Catherine
A.
1965. Nonprofessional Personnel n Physical Therapy.
Physical Therapy
45:122-115.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 273/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 274/375
2
Creating Capitalists:
The Social O rigins of Entre pren eu rsh ip
in Post-Communist Poland
arbara Heyns
New York University
number of papers in this volume deal with career transitions in
post-industria l labor marke ts in a com parative context. The institutional
context how ever is invariably relatively stable and longstand ing. Con
trasting labor markets dominated by stable occupational structures with
those organ ized by stable firms for exam ple as Thom as DiPre te and
Patricia McManus do presum es that the institutions in question operate
within similar markets for labor. Equally if one aims to un de rstand labor
mark et segm entation s a function of union strength an d prevailing policies
toward em ployment and training
s
A rne Kalleberg
does the
presum ption
is that these are set within an established economic stru cture albeit one that
can expand or contract. Variability in education occupational mobility and
earnings are und erstood as a function of the dem and and supply for labor
set by relatively fixed econom ic structures. The capacity of firms to adapt to
changing economic conditions dep end s on manag erial skills and business
cond itions; the poten tial of work force dep end s on hum an capital and on
oppo rtunities for training and mobility. The opp ortun ity structure is
presumed to provide the incentives for investing in human capital and for
individua l labor market b ehavior.
This pap er deals with a different problem and that is how labor markets
form in the first place. In particular I am interested in the gro w th and
expansion of the new private sector and in the development of business
values and entrepren eurial aspirations in post-comm unist countries. Inso
far as market tran sitions entail a transformation of
values
aspirations and
opp ortun ity structures the question becomes how values and aspirations
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 275/375
258
are form ed w he n the op po rtun ity s truc ture is und efined an d fair ly chaotic .
Pre l iminary answers can, I th ink, be found by s tudying s tudent a t t i tudes
and values in pos t -communis t countr ies .
This chapte r w ill pres ent d ata f rom the f irst survey of bu sine ss s tu de nts
in pos t -co m mu nis t Poland. The app roach is com parat ive , in tha t three
different ed uca tion al contexts will be com pa red, bu t the central q ues tion s
concern f irs t , the or igins and relat ive importance of economic values in
shaping career aspirations, and second, the degree to which pre-exist ing
insti tutio ns, such as schooling or the family, influence these a spiratio ns.
Edu cation occupies
difficult a nd amb igu ou s posit ion in pos t-com m un ist
coun tr ies . N ew skil ls , a t t i tudes , and values are need ed, a long with in
creased investments in human capital ; yet the guarantees of jobs and the
securi ty of em ploy me nt pro mise d un der s ta te socia lism ha ve been abrup t ly
term ina ted . Stu den ts m us t be mo bilized to achieve, yet sett ing career goals
in the mid st of chaotic econom ic cha ng e is extrao rdina r i ly diff icult. The
econom ic reforms ha ve a imed to sever the t radi tional l inks betw een edu ca
t ion an d career . The contours of the em erging opp or tu ni ty s t ruc ture are
obscure; the incentives for learning par t icular subject matter are s t i l l un
clear . Knowledge about career possibil i t ies is rather l imited and quickly
ou tda ted . O n the one han d, s tud ents hav e never ha d m ore f reedom to
define their own career goals; at the same t ime, calculating the value of
credent ia ls or the expected re tu rns to educa t ion has never been as prob lem
atic.
Alth ou gh cr it ic ism of r ig id m an po w er pla nn ing as prac t iced by the
former regime is widespread, the connect ions between school ing and
career rew ard s w ere at least clear. The econo mic transfo rma tion m ea ns
fun da m en tal increases in uncer tain ty at al l levels , bu t par t ic ular ly for
s tuden ts .
U nce rtain ties affect all
fields,
but they are ende mic to bus ine ss e duc at ion ,
w he re the purpo ses of t ra ining can be som ew hat obscure , even in the w es t
he
Economist Jun e 4,1994) . O ne mi gh t ask wh eth er i t is rat ional to earn an
unproven credent ia l in bus iness or economics ra ther than becoming an
e n t r e p r e n e u r d i r e c tl y . O n e of t h e m o s t p h e n o m e n a l c h a n g e s i n
pos t-co m mu nis t countr ies has been the rapid deve lopm ent of ne w bus ine ss
schools , course s in econom ics, trade an d co mm erce, as well as ho st of othe r
ne w t r ain ing oppor tun i t i e s fo r bu dd ing en t r epreneur s . P rog ram s r ange
from crash courses last ing a few days or several weeks, to year- long
workshops and seminars , to fully accredited schools offer ing degrees in
bus in ess adminis t ra t ion . They a t trac t s tude nts of a ll ages an d ba ckg rou nd s ,
from yo un g school- leavers with gra nd a mb itions to seaso ned vete rans of
state enterpr ise; f rom opposit ion activis ts eager to dismantle the public
sec tor , to senior managers and planning exper ts who hope to survive the
trans it ion. Classes in account ing and f inance, t rade , com pu ter pro gram
ming , m ana gem ent , m arket ing an d other bus iness ski lls hav e prol ifera ted;
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 276/375
259
the mo st po pu lar car ry the ad de d label in ternat iona l , Eu rope an, or
s imp ly w es ter n . Lan gua ge s tud y, both formal an d informal , i s a lso
exper iencing a renaissance , wi th courses in bus iness Eng l ish leading the
wa y. Rel iable cou nts of the s tud ents invo lved are unava i lable , bu t es t imates
of the nu m be r of cou rses offered in W arsaw alone are in the hu n dr ed s.
1
One
state ente rpr is e ev en offered the old gu ar d of nomenkl tur m a na ge r s
scholarships for executive training seminars as par t of a their severance
package.
W hile it is too ear ly to ha zard an evalu ation of the schools or pr og ram s
that hav e mu sh ro om ed s ince 1989, m uc h less to assess the i r impact on
econ om ic reform, the grow th of bu sine ss ed uca tion is as rem ark ab le as the
cow boy cap ital ism f lourishing center s tage (Grayso n, 1993). Pr ior to
1990,
bus in ess educat io n did no t exis t in com m unis t countr ies . O ne could
s tud y man ag em en t , which mean t the di rec t ion and adm inis t ra t ion of sta te
enterp r ise , or one cou ld s tu dy o r thodo x vers ions of socia l is t econo mics . In
Poland, s ix specialized universi t ies of economics and planning were in
opera t ion, whi le many 'ordinary ' univers i t ies inc luded a depar tment of
econom ics an d m an age m en t . W ith very few except ions , how ever , it w as a ll
bu t imposs ible to learn about pr iva te , inde pen den t , or non-s ta te econom ic
activity.
2
A lek san de r M uller , for exam ple (1991:175) , con clu ded that The
econom ic higher schools have t ra ined use less gradu ates for a non -exis t ing
e c onom y .
Although schools of bus iness are new, they are not ent i re ly unprec
ede nted . N ew educat ion al pro gram s are se ldom cons tructed from w hole
cloth; they d ep en d on exist ing philos op hies , facult ies and facil it ies as surely
as they d isplace the m . If on e env isions establishing a ne w f ield of s tu dy ,
the re are three poss ible strategie s. F irst, on e m ig ht affiliate w ith an e xistin g
univers i ty pr og ram , enlarging a de pa r tm en t wi thin an es tabl ished ins t itu
t ion. Second, one might capture or reform an exist ing specialized school.
Finally, one might establish a new and altogether separate school.
Business education has blossomed in al l three ways in Eastern Europe.
Trad i t iona l management and economics p rograms wi th in the un ive r s i ty
have expanded dramat ica l ly , renegot ia t ing the ir miss ion and credent ia ls ,
wh ile retain ing their conn ections to the un iversi ty . Schools of econo mic
p lan n ing hav e been r eorgan ized , add ing bus iness and comm erce as ma jor
new specializations. Finally, new schools of business administrat ion have
been es tabl ished, typica lly w i th wes tern capi ta l. Th rou gh ou t the region, the
insti tutional l ineages of new and reformed business schools can be traced
to one of these sources.
3
De spite pro fou nd shif ts in philo sop hy and v alue s, inst i tu tions tend to
repro du ce them selves . Despi te new courses and revised cur ricula , schools
tend to recruit an d train s tu de nts w ho share dis t inctiv e— and differ ing—
att i tudes tow ard bus iness . Both the s t ruc ture and organiza t ion are shap ed
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 277/375
260
by preexis t ing ins t i tu t ions ; ne w p rog ram s recrea te the ideologica l tens ions
found in the pro gra m s tha t we nt before . Even an ent i re ly new school , wi th
a radical new m ission, perceive s and re sp on ds to ima ges lodge d in the past .
Con t ra s t ing p rogram s wi th d is s imi la r o r ig ins can p rov id e a b ro ade r und e r
s ta nd ing of the dy na m ics of edu cation al reform. Th e next thr ee sect ions
briefly sketch the three schools s tud ied, em ph asiz ing the major differences
and s imilari t ies .
Specialized Schoo ls of conomics
In
A pril of
1991,
the Pol ish par l iam ent res tored the nam e Szkoia Giowna
Handlowa SGH) to the largest and most renowned school of economics in
Polan d. For for ty-two yea rs , the school ha d been kno w n as Szkoia Giowna
Planowania
Statystyki [Central School of Plan nin g and Stat is tics] , or SGPiS.
With five different facult ies or departments and several branch schools
outs ide Warsaw, the school offered not only s tat is t ics and planning but
m an age m en t , fore ign t rade , and tour ism . Econom ics w as taug ht as an
applied science in the Marxis t-Leninis t mode, a l though the faculty were
familiar with capital is t theory and models . SGPiS was known fo r hav ing
s t rong pol i t ica l connect ions both to the par ty and to s ta te government .
Students at Warsaw Univers i ty, not entirely in jes t , described SGPiS as the
schoo l for police an d spies , ra ther tha n pla nn ers and s tat is t ic ians .
G rad ua tes aspi red to becom e indus t r ia l apparatchiks or go ve rnm en t offi
cials,
and the mos t success ful eventua l ly moved in to h igh pos ts in the
co m m an d econom y. Du ring the 1960s, the school achieved d is t inc t ion and
interna t iona l recogni t ion through the work of economis ts such as Michai
Ka lecki , becom ing kn ow n as the Po l ish Ca m br id ge . Bo th Leszek
Balcerowicz, the architect of the Polish transi t ion, and the former Finance
Minis te r , Je rzy Osia tynski , have worked and s tudied here .
SGH
has made substantial changes in the las t four years , despite the
bu dg eta ry c r ises (Beks iak , Chm ie lecka , and G rze lonska , 1991). D ean s a nd
rectors ret ired or were voted out of office, and the curriculum has been
extens ive ly revised . The dep ar tm enta l s t ruc ture has been co mp le te ly
revamped; s tudents a re now admit ted to the school , ra ther than to a
part icular faculty; their course of s tudy is reviewed by the ful l academic
sena te rathe r tha n by specialized units . A ltho ug h the pr og ra m is s t il l high ly
s t ruc tured , s tudents have cons iderably more choice among courses than
before these reform s. The first f ive sem esters are s ta nd ard ize d, a l t ho ug h
stu de nts are no w allow ed to choose their lecturers ; thereafter , e lect ives an d
tutor ia ls replace com pulsory courses . Enrol lmen ts a t SGH are robus t , and
gr ow ing rapid ly. In 1991-92, there w ere 8 day s tud ents ; in addi t ion , som e
1,600 evening s tudents paid nominal tui t ion fees . SGH recently added an
MBA pr og ra m , m od elle d after Am erican pro gr am s. In 1992, for the f irst
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 278/375
262
time ever, SGH published a course catalogue, listing the programmatic
options that students can pursue (Chmielecka, 1992).
Management and conomics with in the University
The faculty of Management and Economics at Warsaw University is
smaller than SGH but generally considered somewhat more prestigious.
Although Management never achieved the stature accorded more aca
demic disciplines, this program retained a fair amount of intellectual
autonomy und er comm unism because of its affiliation with the univ ers ity.
Course work tend ed to be muc h more theoretical and less applied th an at
the former
SGPiS;
the faculty w ere generally thought to be less political
as we ll.
Like SGH the Management and Economics program has launched
curricular reforms and expanded admissions to meet increased student
dem and ; in 1992,210 students of ma nagem ent and 35 studen ts of econom
ics were enrolled in the undergraduate program, while advanced degree
prog ram s grew dramatically. Postgraduate courses and an MBA degre e, in
collaboration w ith the University of Illinois, are now offered at the Inte rna
tional M anagem ent Centre and School. The progra m s in manag em ent an d
economics always enrolled a large numb er of part-time and evening stu
den ts; since
1989,
how ever, these courses have g rown substantially, desp ite
escalating fees.
Un der the curren t Polish constitution, higher education is free. Ho w
ever, with the passage of the Higher Education Act in 1990, university
program s are perm itted to charge tuition for evening courses or for special
extra-curricular pro gra m s. Fees can also be charged for irreg ular stu
dents, or those returning to school after a certain age. The Management
School, particularly the programs affiliated with the postgraduate Centre,
aspires to become an autonom ous u nit within the un iversity, at least in fiscal
terms. There have been recurring controversies over whether individual
dep artm ents can keep the revenues g enerated by extra-curricular courses.
Compared to SGH dissension am ong faculty over educational reform and
the restructuring of prog ram s seems less prevalent than conflict between
the School and the central university administration over fees and the
disposition of gran ts from interna tional donors . Thus, the W arsaw Univer
sity prog ram in Managem ent a nd Economics represen ts a second site for
business training in post-comm unist countries—an adjunct
to,
and
a
poten
tial subsidy for, university programs.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 279/375
262
Private usin ess Education
The third type of bu sines s school
s
the newest an d, for socialist co un trie s,
the mo st radica l .
The
first p riva te busin ess sch ool
in
Po l a n d ,
the
Prywatna
YJyisza Szkda Businessu Administracji
or the
Priva te Higher School
of
Business and Admin is t ra t ion , was fo u n d e d in 1991.
4
This school is best
descr ibed as the vision of one man, Dr . Tad eusz Kozluk, a Pol ish-American
bus inessman f rom New Jersey and mo st recently , W arsaw. Dr. Koz luk is
justifiably p ro u d of his efforts to establish this school, desp ite en or m ou s red
t ape and prolonged de lays f rom the M inis try of Nat iona l Edu ca t ion . The
school was fina l ly approved in the spr ing of 1991 and the first three
h u n d re d s t u d e n t s a d mi t t e d the follo w ing fall. Like all new e n d e a v o r s in
Po l a n d , the school has faced a n u m b e r of p ro b l e ms in finding app rop ria te
space, recruit ing
staff and
beg inn ing ope ra t ions ;
the
fact that
it has
p ro s
pe red , desp i te these obs tac les , is a tr ibute to the pers is tence of Dr. Kozluk.
A t the t ime of the interview, Dr. Kozluk and the staff had begu n rev iewing
appl ica t ions for the second year ; he expected the next entering class to be
twice as large as that of the first ye ar.
As the na m e sugges ts , Prywatna is thorou ghly pr iv a te . T ui t ion a lone was
two mill ion ztotys per m on th ( 200) in 1 992, close to the average m onth ly
salary
in
the s tate sector; m oreo ver, s tu de nt sch olarship s
or
s t ipe nd s are not
available. In exchange for these s teep fees , s tudents are subjected to a
r i g o ro u s and dem and ing reg ime , encom pass ing not only their cours e w ork
bu t a lso persona l habi ts , groo m ing, outs ide ac t iv i ties, and genera l b ehavio r .
S t u d e n t s are requ i red to at tend all classes, to be punc tua l , o rde r ly , and
w ell-dre ssed; unlike the rest of Warsaw, smok ing s not a l lowed in e i ther the
bui ld ing or the classroo m s. These rules are formalized in a contract that
each enter ing s tudent must s ign wi th the school, a contract that g ov erns
the i r conduct and defines the expecta t ions of the school and staff.
The s tuden t contrac t and the ple thora of rules and regula t ions convey an
explici t m essa ge: Business requ ires discipline; discipline wil l be ta ug ht and
strictly enforced. The objective is to equip s tud en t s wi th the charac ter and
the d i l igence need ed for bus iness , as well as wi th new skills. The sym bol ic
m e a n i n g is clear to s tuden t s and staff alike. Prywatna Wytsza Szkoia
Businessu
Administracji a ims to und o the shoddy w ork hab i ts and lackadai
s ical at t i tu de s inculcated by socia l ism. Po land 's economic pro blem s are to
be unders tood as cul tura l and psychological , rather than systemic; disci
p l ined work hab i t s—and
the
sense
of
responsibil i ty
and
personal d igni ty
tha t resul ts— can prov ide a fresh sta rt for ind iv idua l s and for the econom y.
Each of the three bus iness schools i s un iqu e , a l thou gh the i r s tuden ts tend
to be recru ited from a c o mm o n p o o l of app licants l iving in or near W arsaw.
Each school offers a sim ilar array of bus iness and economics courses , and
each hopes to establish and promote jo int degrees and interna t ional ex-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 280/375
263
change prog ram s with business schools in other countries. Given the
incentives for extra-curricular teaching, even their faculties ove rlap. W ar
saw faculty moonlight extensively; teaching courses in the evening and
part-time at other institutions is a pop ular w ay to augm ent income. The
intellectual ambiance in every school was overwhelmingly pro-reform;
students and faculty in each school endorsed economic reform and wel
com ed the arrival of a market economy w ith enthusiasm . Little if any
nostalgia for the past could be discerned. The extent of actual curricu lum
change at the level of individ ual courses could, of course, be deba ted; at the
same time, the training offered at SGPiS and at the University was consid
ered high quality before 1989. Althou gh the course work differs from
conventional economics and management training in the West, faculty in
each of the schools believes that their program offers high-quality ins truc
tion in business adm inistration. Both faculty an d stud ents believe that
education will be a decisive factor in creating a marke t econom y. Business
schools, even those that formerly trained socialist technocrats, have the
unique mandate of cultivating managerial skills and a vigorous capitalist
ethic.
Along w ith restructuring progra ms , business schools are actively con
structing
new,
post-comm unist reputations.
Prywatna
takes pride
in
being
the first private business school in post-communist Poland, while the two
state schools are absorbed in creating new and reformed images. Consid
ering the question s that m ight be raised abou t their past, it is surely in the
best in terests of the schools to do
so.
These schools are definitely n ot trying
to protect their traditions or rest on their laurels; wisely, perhaps, each
claims to offer a program that is very different than before.
The Research roblem
The sudden flowering of neo-l iberal ism in Poland and other
post-communist countries, even in societies without a significant private
sector, seems to be unprece den ted Kovacs, 1991; Szacki, 1990; Heyns and
Jasinska-Kania,
1993;
Weil, 1993). The dom inan t pa ttern of gov ernance in
the region historically involved authoritarian regimes p residing over largely
agrarian econom ies; the class structure w as pre-industrial, with neither an
established bourgeoisie nor significant m iddle
class.
Commercial activity
w as dom inated by Jews, Germ ans, or other ethnic minorities; landow ners
and the intelligentsia alike viewed business with disda in. Moreover, even
un de r state socialism, the various opposition mo vem ents did not as a rule
advo cate free ma rke ts or liberal institutions. The church, for examp le, w as
certainly anti-communist, but it did not champion laissez faire economic
doctrine s or individu alism . In the political realm, dem ocratic values seem
to hav e antedated the collapse of com munism Weil,
1993),
but the source
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 281/375
264
and significance of neo-liberal economic philosophies are a puzzle .
There is, of course, long stand ing sociological interest in ques tions link
ing values and economic behavior. Hypotheses for studying the values of
post-comm unist stud ents can be loosely derived from Max W eber's thesis
on the P rotes tant ethic and related w ork (Weber, 1930; Co llins, 1980;
Brubaker, 1984). W hile there are precious few Calvinists in Po land, one
would expect that students aspiring to be entrepreneurs would be more
supp ortive of business, more hostile to governm ental regulation and au
thority, and more individualistic; Weberian theory would also predict
future entrep reneurs w ould be more self-reliant, m ore comm itted to hard
work and the acquisitive impu lse, and perhaps mo re concerned abou t the
ethics of commerce. Business students should be less inclined to favor
collective strategies of all sorts, whether aimed at welfare and social
security, at economic p lanning, or at international assistance.
The three business schools chosen seem likely to represent a range of
values and asp irations sufficient to test these hyp othese s. The stud ents
enrolling in SGH the most app lied school with the closest historic ties to
state governm ent, were expected to be the most oriented toward state policy
and toward m anaging state-owned industry. Despite economic and politi
cal reform, SGH students were hypothesized
to be
the most collectivist an d
the least oriented toward laissez faire economics; they were expected to
grav itate to state em ploym ent, as either reformers or planners , rathe r than
to the private sector, and to supp ort state intervention and assistance mo re
readily than students in the other two schools.
In contrast, students choosing to attend Prywatna would be the most
individualistic and the most committed to developing the private sector.
Enrolling in
a
new business school without an established reputation m ight
be considered a costly, high-risk option, wh atever the aspiration s. Corre
spondingly, students at Prywatna were expected to express the strongest
pro-business sentiments,
to be
the most individualistic, and the mos t likely
to anticipate an entrepreneurial future. M anagement stud ents were hy
pothesized to fall between these two extremes in both values and
aspirations.
Value differences among students were assumed to depend on both
socioeconomic background and the choice of school; more importantly,
values were assumed to predict entrepreneurial aspirations among stu
den ts, irrespective of their backgro und s. Ivan Szelenyi and his collabora
tors
1988)
have argued that the origins of entrepreneursh ip unde r co mm u
nism a re predo m inan tly cultural, and can be traced to family value s, rathe r
than to economic ones. Rural entrep reneu rs in Hu ngary expressed values
quite similar to those of a parent or grandparent working under market
cond itions before the Second W orld W ar. These values included au
tonomy and risk taking, resistance to being subordinated to the bureau-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 282/375
265
cratic ord er an d to accepting ascr ibed ra nk s in a hiera rchy , desire to be on e 's
ow n boss , va lue a t tached to ha rd wo rk, an d wil l ingness to de lay c on su m p
tion. (Szelenyi, et al
v
1988: 65) . The auth ors argu e that de sp ite the
for ty-year hia tus , pre-W ar entrep rene ur ia l famil ies had prese rved a dis t inc
t ive belief system, and had socialized their children to these values and
beh avio r . D espite very l imited possibil i ties for inhe r i t ing land or capital
under communism, sna red f ami l ia l va lues exp la ined en t r epreneur sh ip .
Perhaps family relat ions foster neo-liberal philosophies in Poland as well .
These concepts l inking econom ic reform a nd perso nal va lues s t ruc tu red
the inqu iry , bu t they provid ed sens i tiz ing ideas as m uch as formal hy po th
eses. D ur ing the course of the research, a nu m be r of bus ine ss s tud en ts w ere
interv iew ed inform ally, in an effor t to un de rst an d their po ints of view . Th e
majority r idiculed the notion of an economic ethos based on either ethical
or rel igious ideals . In gen eral , the s tu de nts we re skeptical that econ om ic
values ha d any bear in g on econom ic reform or person al aspira t ions . Pro
tes t , they argued, and a decade of pol i t ica l oppos i t ion, had produced
econom ic reform, not Protes tant ism. A l thou gh m an y s tuden ts ha d been
involve d in the informal econom y, they genera lly denied hav ing re la tives
or family me m be rs with conn ections to a f ree m arke t. By their accoun ts ,
entrepreneurship meant nothing more or less than f reedom f rom s ta te
control . Th ey ev en de nie d aspiratio ns to get r ich. The best w ay to m ak e
m on ey in po st-c om m un ist Pola nd wa s to w ork abro ad or for a foreign f irm
base d in W arsaw ; entre pren eur ia l ac t iv i ty w as excess ively risky. A t t i tudes
tow ard the gov ern m en t var ied from cynical to sup po r t ive ; bu t few s tud ents
doubted that poli t ics would dictate the course of economic reform, rather
than individual values or entrepreneurial effor ts .
The Da ta
The survey data to be presen ted are based on ques t ionn aires dis t r ibu ted
to full-time, first-year business students in each of the schools described
abov e. A rran gem ents were m ad e w ith the faculty teaching f i r st -year
classes to permit us to administer forms before class; s tudent interviews
w ere con du cted d ur in g the last full we ek of school in M ay 1992. H en ce, the
sam ple is restr icted to those s tud en ts w ho w ere s t il l active in the last w eek s
of the term and i t probably over represents those who a t tended most
regular ly; in al l , 423 questionnaires were collected, representing vir tually
all of the s tudents present on that day.
5
The stud en ts are no t, of cou rse, a
ran do m sam ple of e i ther W arsaw bus ine ss s tud en ts or the s tud ent s in these
schools; accu rate enro llm ent l is ts w ere s im ply not availab le in any sc hool at
the end of the term. In each sett ing, the data w ere, how eve r , collected in a
similar m an ne r . Stu de nts and faculty were coo perativ e an d quite will ing to
be in terviewed; only a handful of respondents re turned incomplete ques-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 283/375
66
t ion na ires . In each school, a sizeable fraction of the full-time st ud en ts in the
firs t-year classes were interviewed, and the demographic profi le in each
school w as gene ral ly consis tent with other inform ation. The re is l i tt le
reason to believe, therefore, that respondents are a biased sample of the
targe t pop ula t ion . Al th oug h the f indings shou ld be t rea ted wi th caut ion,
the data appear to be representat ive of f irs t-year s tudents in these three
bus in ess prog ram s. Genera l iza t ions to a ll Pol ish s tud en ts , m uc h less to
other po s t -com m unis t coun tr ies , how ever , a re not tenable .
The surv ey da ta a re sum m ariz ed in three sec tions . The next sec t ion
contras ts s tude nts in each of the three schools and sum m arizes the cons t ruc
t ion of m easu res . Stud ent value s wil l be com pa red to those of al l Poles , a t
the outse t of the m ar ke t transition. This section relies on an ex tens ive
bat te ry of ques t ions dev eloped in Britain to m eas ure l ibera l an d autho r i ta r
ian at t i tudes cross-cultural ly (Heath, 1991; Biatecki an d M ach , 1992). T he n,
I examine the de terminants of s tudent va lues in Poland, and compare the
differences b etw ee n schoo ls. The objectives in this section are tw ofo ld; first,
to un de rs ta nd recrui tm ent pa t te rns am ong schools, and second, to com pare
schoo ls in term s of valu es and s tu de nt ba ck gro un d. The next section
eva lua tes s tuden t career aspi ra t ions , and en dea vor s to predic t w ho aspires
to ow nin g their ow n business in Po land . The final section co m pa res m ale
and female s tudents in terms of values and aspirat ions.
T h e S t u d e n t s
Polish s tudents entering business school in 1991-2 were truly pioneers .
These s tude nts w ere the f irst c lass to be adm it ted to bus in ess an d m an ag e
ment s ince the beginning of the market t ransi t ion and they wil l be among
the f irst gra du ates of these new ly-formed— or reform ed— prog ram s. De-
mographically, the business s tudents resemble other f irs t-year universi ty
s tud ents . T he major ity (54%) we re male . N inety percent of the resp on den ts
w ere be tw een 18 an d 20 years of age ; two -thi rds w ere exact ly tw enty . Less
tha n ten percent had any pre viou s work o r bus iness exper ience . Ov er hal f
(51%) of the s tu de nts ha d l ived in War saw their entire
lives.
O n e in six (17%)
ha d l ived abroad for three or m ore m on ths du r in g the last three years ; m ore
tha n a third (38%) expected to travel abroa d in the nex t three ye ars . The se
data suggest that the business s tudents are a fair ly sophist icated and
cosmopol i t an g roup .
In t e rms o f soc ioeconomic background , the s tuden t s a re c lea r ly
ad va nta ge d. Half of their m oth ers and s ixty per cen t of their fathers ha d at
least some universi ty education. Their family incomes were by Polish
sta nd ar ds qu ite hig h; over half (54%) rep orte d th at per capita family inco m e
w as ov er 2 m il l ion
zhtys
pe r m on th, or abo ut 200. (In Po land as a w ho le,
the ave rag e pe r capita family inc om e in 1991 w as 766,000
zhtys
per m onth ,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 284/375
267
or less tha n eighty dollars .) Forty perc ent had at least on e pa ren t e m pl oy ed
in the pr iva te sector . These f igures imp ly tha t W arsaw bu s iness s tu den ts a re
a re la t ive ly pr iv i leged group, even among univers i ty s tudents .
6
In severa l im po rtan t respec ts the ba ck gro un ds of the s tuden ts in the th ree
schools differ , most noticeably between the private and s tate sectors of
educa t ion . S tud en t s a t t end ing Prywatna l ive in families w ith sub stantia l ly
higher incomes than the s tudents in e i ther the Management program a t
W arsaw Unive rs i ty or Szkoia Gtowna Handloiva SGH ). In add i t ion , pr iva te
school s tud en ts were mu ch m ore l ike ly to hav e se l f-employed par ents
w or kin g in the pri va te sector. C ons ider ing the tui t ion , these differences are
not te r r ib ly surpr is ing . W e wi l l re turn to th is i s sue w he n w e discuss s tud ent
recru itm ent an d differences b etw een schools ; f irs t, it is nece ssary to review
the measures of economic and poli t ical values .
Political and Economic Values M easuring Liberalism
Measuring poli t ical and economic values in post-social is t countries
po ses difficulties . The pro ble m is twofold: research on poli t ical an d
econom ic va lues was no t wel l -dev eloped in com m unis t coun tr ies , and few
s tud ie s p roduced re levan t measures tha t cou ld be compared ove r t ime .
Exis t ing va lue sca les wi th known proper t ies inc lude i tems tha t may be
i r re levant to countr ies in t rans i t ion; moreover , such measures may be
prob lem at ic for c ross -cul tura l com parison. O ne is , ho w eve r , forced to
choose be tween deve lop ing new and o r ig ina l measures wi th unknown
proper t ies , or adapt ing i tems tha t were in i t ia l ly developed and tes ted in
othe r co untr ies (Co nve rse, 1964; Rok each, 1979).
For th is s tud y, how ever , only one opt ion m ad e sense . A ba t t e ry of
eighte en i tem s con stru cted to m ea su re core va lue s of l iberal ism in Bri tish
e lectora l s tudies had been t rans la ted and a dm inis te red to a ran do m sam ple
of Polish adu lts in 19917 T he English research con cep tualize d l iberal ism as
an endur ing se t of va lues , charac ter ized by three d is t inc t ive d imens ions .
Econom ic va lues a re def ined a long a con t inu um repre sent in g the deg ree of
sup po r t fo r laissez-faire econom ics op po sed to soc ia l is t a r rang em ents ; pol i t i
ca l va lue s a re m ea sur ed on a scale tha t var ied f rom l iber ta r ian to auth or i ta r
ian. The third dimension represents individualis t versus collect ivis t a t t i
tud es . These i tems hav e sub stantial face validi ty, an d they ha ve bee n s ho w n
to yield s table and rel iable measures of poli t ical and economic values in
other na t iona l survey s . M oreover , the i tems we re be ing used in para l le l
ana lyse s ba sed on seve ral othe r subp op ula t ion s (Biatecki an d M ach, 1992).
Hence , comparisons were poss ib le among di f fe rent groups of Pol ish re
s p o n d e n t s .
All i tems are measured on a four-point scale , from strongly agree to
strong ly disag ree. A confirm atory factor analysis w as per form ed o n the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 285/375
268
items in each of the domains for the national and s tudent samples sepa
rately . Each set of item s yields a single un de rly in g factor scale, w ith the
princ iple co m po ne nts acco untin g for be tw een .36 an d .43 pe rcen t of their
respective co m m on va r iances. The factor loading s range d from .46 to .74
across i tems; the relat ive contr ibutio n of each i tem to the com po site scales
wa s s imilar in both sam ples , a l thou gh th e bus in ess s tud ents scored high er
than the nat ional sample on vir tua lly a l l of the individ ual i tems m eas ur in g
l issez f ire
or l iber tar ian att i tudes.
8
Th e reliabilities calc ulate d for th e
s tud en t sam ple were , how ever , as h igh as those for the po pu la t ion , desp i te
the impl ied ceiling effects. The alp ha reliabilities calcu lated for the s tu d en t
sam ple an d the i tems inc luded are given below .
1) La issez Faire - Socialist scale (7 Item s, A lp h a = ,787).
9
2) Liber tar ian - A uth orita r ian scale (6 I tem s, A lph a = .592)
10
3) Collectivist - Ind ivi du alis t scale (5 Item s, A lph a = .571)."
These rel iabil i t ies compare favorably with the Brit ish results ; the s imi
lar i t ies with the national sample suggest that these three composite mea
sures capture consis tent differences in values among the Polish respon
den t s . Th e /fl 'ssez-/(jfre-socialist scale can be in te rp re te d as eit he r p ro -b us in es s
or as anti-socialist; libertarian-authoritarian values reflect respect for indi
vidual r ights versus support for law, order , and tradit ional authority. The
measure of individualism-collectivism seems to tap a sense of personal
respons ibi l i ty as wel l as puni t ive a t t i tudes towards welfare and depen
denc y. Al th oug h these i tems were not developed to assess va lues or va lue
cha ng e for
po pu latio n in transit ion, the three scales app ear to disc r im inate
bet w een gro up s wi th di fferent econom ic an d pol it ica l va lues a t the o utse t
of the econom ic transit ion.
How Liberal Are the Polish Business Students
The three sca les w ere cons tructed to m easu re "co re" va lue s , an d hence
relatively s table and enduring trai ts . Since their explanatory power, even
as descr iptiv e m eas ure s is no t kn ow n, the f irst task is to locate the bus ines s
students in terms of economic and poli t ical values on each of the three
dimensions measured with reference to the nat ional popula t ion survey.
The bus in ess s tude nts are , obviou s ly , an especia lly adva ntag ed gro up . In
terms of age , income, educat ional background, and urban res idence , the
s tud en t samp le is qui te se lec t. Age , educat ion, and urb an res iden ce h ave
each be en sh ow n to inf luence v alues; i t w ou ld be diff icult to im agi ne that
these and related factors are uncorrelated with at t i tudes and values.
In ord er to com pare bus iness s tude nts wi th the nat iona l po pu la t ion , each
of the three value me asu res w as regressed on the full range of back gro un d
factors com m on to bo th surveys . The ind ep en de nt var iables inc luded age ,
sex, urb an location, farm origins, the edu catio n of the ho use ho ld he ad and
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 286/375
269
the spou se , the sec tor of em plo ym ent and the to ta l family inco me repo r ted
by the res po nd en t. Since the major i ty of s tu de nts l ived at ho m e, their
parents educat ion and family income were used ra ther than the ir own.
Three dummy var iables were a lso added, indica t ing whether or not the
respondent had l ived in Warsaw for most of their l ife , whether or not
any on e in the ho use ho ld received a s ta te pen s ion, and w heth er any one w as
une m ploye d .
T he
laissez-faire,
liber tar ian, an d ind ividu alism scales der iv ed from the
factor analy sis w ere regressed on these ba ck gro un d factors . In each
equat ion, a dummy var iable for s tudent s ta tus was inc luded to tes t the
overa l l d i f ferences between the two samples ; then three separa te dummy
variables, one for each school, were substituted, in order to test for differ
ences be tw een schools . I a lso exam ined interac t ions betw een school type
and bac kgro und s ta tus . S tuden t s ta tus on bo th
laissez faire
an d l iber tar ian
va lues s hig hly s ignif icant, un iqu ely acco un ting for be tw ee n 6-8 pe rce nt of
the total var ianc e. O n the individualis t-collectivis t dim en sio n, how ev er ,
de m og rap hic an d economic backg rou nd fac tors are la rgely unre la ted to the
value scores , an d the dif ferences betwe en s tu den ts and the genera l po pu la
tion are not significant.
Table 12.1 pre sen ts the ordin ary least squ ares regression of the va lues
scales on socia l bac kg rou nd for the m erged n at iona l and s tu de nt sam ples .
As the regres sions clear ly indicate, the s tud en ts differ dra m atically f rom the
average Polish ci t izen and, not surpr is ingly, especially on the dimension
m eas ur in g econom ic values . M oreover , these di fferences pers is t ir respec
tive of th e full set of con tro ls. Laissez faire value s tend to be the m os t l inke d
to socioeconom ic factors, while l iber tar ian v alue s tend to be a function of
age an d urba n res idence . W ith and w ithou t controls , the bus ine ss s tu den ts
are near ly a full s tandard deviation more
laissez-faire
than the Polish
pop ula t io n and appro xim ate ly half a s tan da rd devia t ion m ore l iber tar
ian. Even w he n the pop ula t ion sam ple is res t ric ted to those wi th so m e
high er educ at ion or to those wh o are un de r the age of 35, s tud ent s ta tus
increases the var ianc e exp laine d in econ om ic an d poli t ical valu es by at least
three percen t. In con trast , with respect to ind ividu alis t - collectivis t valu es,
the student scores are not significantly different from those of the national
sample .
The regress ions prov ide sup po r t for the propo s i t ion tha t these m easu res
of l iberalism yield value configurations for post-communist countr ies
resem blin g those observ ed in the w est . At the sam e t ime , the s tren gth of
their associations with tradit ional social factors , such as class or demo
graph ic bac kg rou nd , tend s to be lower than in Europe or the Un i ted S ta tes .
There is evidence tha t economic background plays a role in economic
value s ; ha vin g e i ther paren t own ing a pr iva te bus in ess is second o nly to
parenta l educat ion as a fac tor predic t ing how s t rongly laissez-faire values
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 287/375
TABLE 12.1 Un stan dard ized
Values Regressed
including the
Independent Var iables
Sex
Age
Education Wife
Educat ion Husband
Farm Background
Sector of Emp loymen t
Pr iva te Owner
Family Income
Pension
Unemployed
Size of Town
Warsaw Resident
Student
Intercept
R
2
Student SGH
Ma na ge me n t
Prywatna
R
2
b
-.013
-.002
.095
.055
.148
.018
.233
.019
.013
-.245
-.002
.007
.906
-1.064
.3607
.866
1.013
.950
.3651
Regre ssion Coefficients of Laissez-faire Libe rtarian an d Individ ualist ic
on Socia l and Demographic Background Factors
nat ional random sa
Laissez-faire
s.e.
.047
.002
.016
.015
.092
.034
.091
.010
.073
.183
.000
.318
.083
.134
.083
.097
.135
t
.70
1.05
5.79
2.90
1.61
.35
2.56
1.98
.86
-1.34
-.56
.80
10.96
-7.89
9.94
9.32
6.63
for the
merged Pol i s r
s a m p l e
mple [N=915] and the bus iness school s tudent s [N=423] .
b
-.035
-.013
.031
-.010
-.068
.011
-.045
-.016
.092
.039
.153
-.003
.425
.302
.2126
.465
.427
.315
.2214
Libertarian
s.e.
.050
.002
.017
.015
.101
.059
.100
.011
.080
.201
.017
-.098
.051
.148
.097
.119
.148
t
-.48
-6.56
1.86
-.49
-.67
.85
-.65
-1.16
1.15
.20
3.48
.97
4.78
2.04
4.81
3.59
2.13
b
-.057
.003
-.032
.036
-.011
-.001
.083
.027
-.184
-.151
-.081
.043
-.107
.037
.0306
-.177
-.113
.194
.0395
Individual i sm
s.e.
.056
.002
.019
.016
.334
.034
.113
.012
.088
.201
1.651
.971
1.06
.167
.109
.135
.167
t
-1.02
1.47
-1.67
2.05
-.78
-.97
.74
2.13
-3.10
-.62
-1.09
.33
-.29
.22
-1.62
-.84
1.16
N9
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 288/375
271
are exp resse d. Libe rtarianism , in contra st , is largely a function of age a nd
urb an res idence , w hich do not d iscriminate am on g the s tud ents . As o ther
analys ts hav e found , soc ioeconomic s ta tus predic ts poli t ica l a t t i tud es and
values in pos t -c om m unis t countr ies ra ther we akly . This i s typica l ly in ter
preted as indicating a dis juncture between social s tructure and values or as
indica ting tha t class interests hav e no t yet crystal lized (A dam ski, 1993).
Th e final th ree r o w s give n in Table 12.1 rep re sen t the net effects of each
of the schools , with the ful l set of background and demographic factors
contro lled. These coefficients are the un sta nd ar di ze d v alues of the three
dummy var iables when school type is subs t i tu ted for the s ingle s tudent
s tatu s variab le. Th e coefficients represen t dev iat ion s from the om itted
catego ry, in this case the natio nal pop ula t ion sam ple , on each scale . The
differences be tw ee n schools are no t large; net of ba ck gro un d variables they
are significant only for the scale m ea sur ing ind ivid ual ism . The scores do ,
how eve r , cor resp ond to the hyp othe s ized order ing be tw een schools. The
priv ate bu sin ess school is the m ost individu alis t ic an d the least l ibertarian ;
SCH is the least
laissez-faire
the most l ibertarian and the most collect ivis t .
On a l l three sca les , the Management program a t Warsaw Univers i ty fa l l s
between the other two schools .
The nex t sec t ion explores the re la t ionships be tw een these cor e va lues
and o ther charac teris t ics wi th in the bus in ess s tudent sam ple . I t m us t b e
em ph as ize d, ho we ver , tha t in te rms of econom ic and pol i t ica l va lu es , the
Warsaw bus iness s tudents a re subs tant ia l ly more l ibera l than the average
Pol ish c i t izen, even when compared to those wi th s imi lar demographic
charac ter is t ics an d comp arable economic and cul tura l resources . The
evid enc e that the thr ee bu sines s schools hav e a dis t inct ive pat ter n of valu es
is m uc h mo re tentat ive. A ltho ug h the differences be tw een schools are in the
pred ic ted d i rec t ion, the ir m agn i tud e is not la rge . W hen d em og rap hic
fac tors and soc ioeconomic background of s tudents a re control led , va lue
differences between schools are insignificant .
Va lues and Student ecruitment
Despi te the en orm ou s dem an d for bus ine ss edu cat ion tha t now exis ts , to
so m e exten t these thre e bu sin ess schools com pete for clientele . Each of the
schools tends to recrui t s tudents f rom the same ca tchment a rea , which
inc ludes Wa rsaw an d the sur rou nd ing region. M oreover , the schools d i ffer
in cost , in expectat ions for s tudents , and in their ass igned posit ions under
the prev ious regim e. Al th oug h only the pr iva te school i s or expects to
rem ain com ple te ly dep en de nt on tu i t ion , a t t rac t ing s tuden ts is increas ingly
impor tan t th roughout Po l i sh educa t ion .
In spite of reform, the two established s tate schools were expected to
recruit a s tu de nt bo dy that resem bled their his toric cons ti tuencies . The ne w
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 289/375
272
pri va te school in contrast m us t secure a niche in the m ark etp lace largely
by co m pe ting for the cl ientele of s tate schools . W e expe cted
Prywatna
to
a t t rac t s tud ents w i th a s t rong com m itme nt to the new p r iva te sec tor w ho
would be wil l ing and able to pay high tui t ion in a somewhat r isky set t ing.
Sta te schools h ow ever despi te lower tu i t ion and nu m er ou s reform s no
longer have secure and well-defined channels for ei ther recruitment or
plac em en t. H enc e the characteris t ics of s tu de nts at each school can be
tho ug ht of as indica tors of both cont inui ty a nd ch ange be tw een schools .
The observed va lue d if ferences be twe en schools sup po rt those hy po th
esize d in ter m s of the relative m ag n itu de of effects eve n if the differences
in abso lute s ize are no t s ignificant . The first quest io n that mu st be ask ed
ho w eve r i s w heth er these d is t inc t ive va lue conf igura t ions sh ould be con
sidered an intrinsic aspect of each school or a consequence of differential
recrui tm ent pa t te rns . A l thou gh the re i s l it tle var iance be tw een schoo ls in
age paren ta l educa t ion or sex com pos i t ion as we hav e seen the schools do
differ in the socioeconomic s tatu s of their s tude nts in consis tent wa ys . In
this sect ion we wil l pursue the quest ion of whether socioeconomic differ
ences between schools account for expressed value differences .
Tw o dis t inct hy po the ses gu ide d this wo rk. Firs t w e expe cted differ
ences be tw een the pr iva te an d publ ic sec tor in bus iness educ at ion; second
we expected differences within the public sector would reflect whether
s tudents were recrui ted f rom Warsaw or f rom outs ide . Parenta l employ
ment should be reflected in school choice because private sector parents
m ay prefer pr iva te educa t ion and because pr iva te schools a re cons iderably
m or e expen s ive than the i r publ ic sec tor cou nterp ar ts ; s tu de nts enrol l ing in
Prywatna
as we have seen are from families with substantial ly more
inco m e than in ei ther of the other tw o schools . M oreov er in this f irs t year
of opera t ion Prywatna adm itted s tu de nts quite la te in the year after ex am s
ha d been com plete d by the other two facult ies . We w ere interested in
comparing s tudents who chose to a t tend pr iva te bus iness school over the
state al ternatives in term s of their valu es and objectives .
Historical ly the m ost s ignificant difference bet w een th e tw o s tate schools
is a con sequ enc e of special izat ion an d the deg ree of em ph asis on practical
as op po sed to theoret ical econom ics. Both the trainin g an d the crede ntials
offered by SGPiS were subs tant ia l ly m ore vocat ional ly-or iented th an those
of the Univers i ty prog ram . The appl icant pool w as a lso s igni ficant ly m or e
diverse and inc luded s tu den ts f rom sm al l tow ns an d rura l a reas in Polan d.
Stu den ts gro win g up outs ide Warsaw tend to be m ore provinc ia l than th ose
w ho at ten de d seco nda ry schools in the ci ty. W ithin the pu blic sector w e
expected bus iness s tuden ts to hold d if ferent va lues de pe nd ing on w heth er
or no t they grew u p in W arsa w . If va lue differences be tw een scho ols can be
exp lained ful ly by these three characteris t ics family inc om e sector of
em plo ym en t an d family res iden ce i t s trong ly sugg ests that valu e differ-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 290/375
73
TABLE 12.2 Logistic Regression Equations Predicting School Choice by Student
Background, Polish Business Stud ents, 1992.
D e p e n d e n t V a r i a b l e s
Independen t Pr iva t e
Variables Coefficient
App l i ed e l sewhere
Passed Exams
Family income
Self-Employed
Warsaw Resident
Values
Laissez-Faire
Libertarian
Individual ist ic
Cons t an t
-2 Log Likel ihood
Model Chi -Square
Deg rees of freed om
Number of cases
Goodness of Fi t
3.95
-1.85
.17
1.49
-.07
-.04
-.04
.37
-3.99
vs. State School
s.e.)
.57
.59
.07
.41
.38
.27
.23
.20
.53
201.19
111.79***
8
423
427.18, p=.016)
Sig-
.000
.002
.014
.000
.851
.875
.847
.069
.000
SGH vs. M a n a g e m e n t
Coefficient s.e.)
-1.48
.12
-.17
.13
-.51
-.17
.11
-.09
1.93
.80
.86
.05
.35
.26
.18
.15
.13
.32
370.00
35.82**
8
362
328.41, p=.251)
Sig-
.065
.893
.003
.713
.051
.325
.463
.494
.000
ences between schools should be attributed to recruitment, rather than
socialization. Although a particular business school may reinforce values ,
differences that are du e to such background factors imply tha t prior family
influences, rather than business school experience, are critical for the
emergence of new economic and political values.
Table 12.2 presents two logistic regressions designed to test these hy
potheses. The dependen t variables in each case are the log odds of enrolling
in a particu lar school.
The first equation p redicts priva te sector enrollm ent com pared to the two
state schools, wh ile the second compares only the two schools in the public
sector,
S H
and the University of W arsaw M anagem ent program . In each
equation, Warsaw residence, family income and sector of em ploym ent are
entered, with each of the three dimensions of liberalism previously dis
cussed. Two additional dum m y variables assessing recruitmen t are in
cluded: w hether or not the studen t had applied to another business school
an d, if
so,
whe ther the student had passed the exams for that school. Only
a minority of students (20 ) had applied to a second school, and two-thirds
of these stated that they had either passed the exams or that none were
requ ired. However, these variables account for a large pa rt of the difference
between the private and public sectors in business education.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 291/375
74
The coeff icients est imated in the f irs t equation address the question of
differences between sectors, or in this case differences between Prywatna
and the two state-supported schools . The results imply that both sector of
paren ta l emp loym ent an d family incom e are s t rongly and pos i t ive ly re la ted
to the prob abil i ty of enroll ing in the pr ivate school, wh ile Wa rsaw reside nce
is not. Stu den ts cho osing Prywatna we re the m os t likely to ha ve ap plie d to
at least on e other busin ess school (65 ha d d on e so) , an d the least likely to
have passed exams .
Prywatna
w as ap pro ve d quite late in the sp r ing of 1991,
and , s ince appl icants could not kno w in adva nce w heth er the school w ou ld
be ap pro ve d, mo st prospect ive s tuden ts had a l ready appl ied for adm iss ion
to other schools . But to some extent, Prywatna seem s to ha ve se rved as a
back-up opt ion for s tudents who might be denied admiss ion; a t the same
t ime ,
ne arly half (47 ) of those atte nd in g Prywatna who had sought
admiss ion e lsewhere sa id tha t they had passed the exams; presumably,
these s tuden ts had chosen Prywatna ov er the establish ed s tate schoo ls . A n
im po rtan t function of pr iv ate sector bus ines s schools , ho w ev er , seem s to be
allowing relat ively advantaged students access to training and credentials ,
de spi te be ing denied adm ission to pub lic sector schools. The major factor
determining enrol l ing in a pr iva te school , however , which holds equal ly
am on g those w ho did an d did not appl y to ano ther school, is family inco me .
When socioeconomic factors are controlled, differences in values are
unifo rmly insignif icant be tw een the pr iv ate and the pu blic sector . Indi
vidualism is the only value dimension that approaches s ignif icance, and
even this difference is small when background differences are controlled.
The coeff icients in Table 12.2 include the three value measures, al though
they are correlated; enter ing them separately, however , does not s ignif i
cant ly im pro ve the ir explanatory pow er . The interac t ions betw een back
ground and values , a l though somewhat la rger than the main ef fec ts , a re
also insignificant. These results imply that vir tually all of the observ ed valu e
differences between the public and the pr ivate sector in these data can be
expla ined by dis t inc t ive recrui tment pa t terns be tw een sec tors .
The second equat ion compares recrui tment be tween the two publ ic
schools , omitt ing the Prywatna stu de nts . Socioeconomic s tatu s is less
important in predicting enrollment within the public sector than i t is in
dist inguishing public f rom private sector enrollments . Family income and
W arsaw res idence are both s ignif icantly high er for M ana gem ent s tud en ts ,
ho w ev er . On ce these tw o factors are controlled, the differences in the
liberalism scales between
SGH
and the Man agem ent School are smal l a nd
insignif icant. Although
SGH
s tude n t s t end t obe bo th m o r e c o l l e c t i v i s t a nd
anti-authoritar ian than s tudents at the Management School, these differ
ences appear to depend on family background. SGH students , as noted
prev iou sly, are the most l ikely to be recruited from ou tside W arsaw ;
corre spo nd ing ly, family inco me s are low er . On ce inco me and place of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 292/375
75
residence are controlled, how ever, value differences between the two state
schools are insignificant. The two state sector schools still recruit and enro ll
students that resemble their former constituencies to a striking
degree.
The
second equation was also estimated by entering each value dimension
separately ; these coefficients were slightly higher, of
course,
but still insig
nificant in every case. Additional background variables, including those
introduced in Table 12.1, do not alter these conclusions.
To summ arize, economic and political values am ong the W arsaw busi
ness stu dents are related, in the expected direction, to both family income
and to hav ing a self-employed paren t in the private sector. Value differ
ences among the schools, although consistent with predictions, are small
and largely insignificant. Differences between schools appear to be a
consequence of the dem ograp hic and socioeconomic factors that determ ine
selective recruitment and enrollment. Variation in liberal values appear s to
be largely a consequence of residence, private sector experience of parents,
and family income. Paren t's education, which is uniformly high in all three
schools, does not pred ict liberal values independen tly of pa ren t's economic
status.
The residual differences in the liberal values measured here, do not
justify assuming a strong causal role for business education in shaping
values. Perhaps one year of business schooling is too short a time to
influence economic and political values . Or pe rhaps stude nts acquire
laissez-faire
libertarian and individualistic values from their families di
rectly, prior to entry . In any event, once socioeconomic background is
controlled, liberal values are not significantly different between schools,
irrespective of private or public sector control or the institutional legacy.
For the two state schools, demographic patterns appear to be consistent
with those that prevailed before the transition
began,
suggesting that these
two schools continue to recruit from the same applicant pools, despite
educational reform. Prywatna in contrast, seems to be carving out a
specialized niche, as a consequence of the high demand for business
schooling and growth in the number of families who can afford private
education.
Student spirations
W arsaw business school students, even more than their pa rents, aspire
to work in the fledgling priva te sector The most popular career destination
is a private firm in Poland (39 ), followed closely by owning one 's own
business (38 ), or work ing for a foreign firm, (35 ). In every school, more
studen ts asp ired to own their own business than to work for a foreign firm
or a joint ven ture in Poland . Moreover, employm ent in state firms w as
overwhelmingly rejected, despite the fact that this had been the dominant
career track unde r com mun ism. Less than ten percent (9.4 ) of the students
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 293/375
276
aspire to w or k in a s tate firm in Pola nd , al thou gh close to on e-third con cede
that reforming s ta te enterpr ises wou ld be chal lenging. W hen asked w ha t
they con sidered the greatest cha lleng e confronting Polish bu sine ss , 40.5
pe rcen t of the s tu de nts said s tar t ing a ne w f irm, 30.4 perc ent said restruc
tur in g s tate ente rpr ise , an d 9.6 perce nt said pr iv atiza tion . O ne ou t of
f ive s tudents aspires to work abroad, while research in a universi ty or
scientific institute (4%) was the least popular career choice. The major
reason s tude nts chose to a t tend bu s iness school, by a cons iderable m argin ,
w as tha t it pro vid ed a chance to ma ke m on ey , wi th
58
percen t m ent ion ing
this goal . The second mo st impo r tant reason wa s som ew hat m ore patr io t ic ;
29 perc ent s tar ted this cou rse of s tu dy because the Polish eco nom y n ee ds
educa ted bus inessmen.
1 2
The career aspirations of s tudents differ more than their economic and
poli t ical values, al though only
Prywatna
students differ once gender and
family ba ckg roun d are control led. In every school , aspir ing ent repr ene urs
are more l ikely to be male, and to have fathers or mothers who are
se l f -employed and earning more income than average . Future entrepre
ne ur s differ in term s of valu es as well . Tho se w h o aspire to s tar t their o w n
bus iness es score s ignif icantly hig her on
laissez faire
att i tudes as well .
By w ay of sum m ary , Table 12 3 presen ts the logist ic regression of ho pi ng
to s tar t one ' s ow n bus iness on se lec ted backg roun d v ar iables and econom ic
values . The bac kg rou nd var iables inc luded in the equat ion are gen der an d
having at least one self-employed parent; family income is largely redun
da nt wh en pr iva te sector em ploy m ent and the school a t tend ed is inc luded.
Tw o addi t iona l me asures of s tude n t mot iva t ion a r e inc luded : the nu m be r
of occupat ions chosen by the respondent and two dummy var iables , one
indica t ing tha t the pr im ary reason the s tud ent chose bus iness school was to
m ak e m one y and the other indica t ing tha t s tar t ing one 's ow n bus ines s w as
the greatest challenge for Polish bu sines s . A ltho ug h these var iab les are
quite highly related to the dependent var iable, they improve the f i t of the
logistic m od el with ou t redu cin g the s ize or s ignif icance of the i nd ep en de nt
var iab les of interest . The school at ten ded w as inclu ded as a categoric
var iable with two degrees of f reedom, and coded as deviations f rom the
overall m ean . Finally, scores on laissez faire valu es an d o n a var iable
measur ing the extent to which the respondent be l ieved tha t the economy
shou ld be comp lete ly unre gula ted w ere inc luded . Tes ts we re con du cted for
f irs t- and second-order interaction terms, as well as a number of other
var iables; these were uniformly insignif icant and are excluded from the
final model.
Th e final col um n of Table 12.3 pre se nts th e antilo gs of the logistic
coefficients, allowing one to see their relative strength as additive effects.
The valu es imp ly, for exam ple, that m ales we re m ore tha n tw ice as l ikely to
aspire to owning their own businesses, s ince e
72
= 2.05. This effect, with
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 294/375
77
TABLE 12.3 Logistic Regression of Own ership A spirations on
Student Background and Economic Values, 1992
Independent
Variables Coefficient (s.e.) p_ Exp(B)
Sex
Self-Employed
Business School
M a n a g e m e n t
S G H
Laissez-Faire
Chal lenge - Own
Money
Number o f Ca ree rs
No Regula t ion
.720
.589
-.669
-.137
.296
.663
.406
.839
.197
.231)
.264)
.188)
.160)
.154)
.224)
.226)
.247)
.089)
.0019
.0255
.0004
.3914
.0542
.0031
.0728
.0007
.0280
2.05
1.80
.51
1.15
1.35
1.94
1.50
2.31
.82
Constant -1.616 .438)
-2 Log Likelihood 490.190, with 414 deg rees of freedom
Model Chi-Square 15.958 p = .0003
Goodness of Fit 372.789 (p = .4063)
oth er var iables controlled , is the largest substan tive difference in the m od el,
s ince to som e extent, the nu m be r of careers chosen necessar i ly increase s the
log od ds of w an t ing to ow n one s own bus ine ss . W e wil l re turn to th is point .
The school attended exerts a smaller influence, but as the coefficients
indica te , Prywatna s tudents are the most l ikely to aspire to ownership.
Private busin ess schoo ls , al th ou gh st i ll very small , seem bo th to enroll an d
train a disproportionate share of the pr ivate sector .
The f inal mo del correctly pre dicts 70 perc ent of the s tu de nt s choices
betw een bus iness ow ners hip and a ll o ther careers. The tw o back gro un d
measures , sec tor of parenta l employment and gender , expla in most of the
differences between Prywatna a nd t he M a na ge m e n t p r og r a m ; SGH s tu
d e n t s ,
however , remain the least l ikely to aspire to own their own busi
nesses . The values predic ted to expla in prop r ie torsh ip aspira t ions are a lso
s ignif icant ; aspir ing owners are among the most
laissez faire
and the most
l ikely to think the eco no m y require s no regula tion at al l. V alues co ntr ib ute
to the explanat ion of ownership aspira t ions independent ly of background
or other at t i tudes; potential owners are also the most l ikely to regard
star t ing a business as the biggest challenge, and the most l ikely to have
entered b us ine ss school in orde r to m ake money- Al th ou gh these resul ts a re
not too surpr is ing, they suppor t the conclus ion tha t economic values
dist inguish different career tracks among these future Polish businessmen,
even when s tudent backgrounds , aspira t ions , and choice of school are
con trolled. M oreo ver , they sugge st that one of the pr i m ary functions serve d
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 295/375
78
by the private sector in business education is to provide training and
cred entials for priv ate sector en tre pre ne ur s . A ltho ug h aspirat io ns differ by
socia l background and by choice of school , Warsaw bus iness s tudents
aspi re to pr iva te sec tor jobs and ent repreneurship in la rge numbers . The
single factor that seems to reduce the desire to own one 's own business is
gen der . An d to tha t we now turn .
Gen der V alues and Proprietorship
Th rou gh ou t Eas t Cen tra l Euro pe , one of the mo s t d isconcer t ing t rend s
has been the inc rease o f gender -based inequa l i t i e s (E inhorn , 1993 ;
M ogd aham , 1993 ;Mul le r and Funk ,
1992;
Hause r , Hey ns , and M ansbr idg e ,
1992;
Rueschem eyer , 1994). W om en tend to be ove rrepres ented am on g the
un em plo ye d and und errep resen ted in the ne w pol it ica l leader ship . Al
though mining and heavy indus t ry were genera l ly regarded as the mos t
problematic economic sectors , the f irs t f i rms to be closed were in the
industries dominated by women, such as texti les or small e lectrical appli
ances . Even in those sectors of the econ om y in grea t de m an d, suc h as the
services , women have tended to remain in the impover ished s ta te sec tor ,
w hile m ales priv atiz e their m edic al practices , law offices, an d busin ess
services of all sorts . Ch ild care, m ater na l leave, an d re pr od uc tive r ig hts of
all sorts are defined as forms of socialist paternalism, too costly to be
re ta ined, whi le un em plo ym ent benefi ts and re t ra in ing for m en are cons id
ered urg ent pr ior i ties . Informal ly , there i s eno rm ou s pressur e on wo m en to
ret ur n to the ho m e, to accept the choice of ful l- t ime m ot he rh oo d that was
denied them under s ta te soc ia l i sm, whi le men are encouraged to become
entr epr ene urs and bus inessm en. The m arket t rans i t ion in Eas tern Eur op e
seem s to ha ve redefined the concept of priv ate along g en de r l ines: for
w om en , pr iv a te m ean s the hom e and a re tu rn to dom es t ic i ty , wh i le for
m en i t m ea ns ent r epr ene ursh ip and econom ic ac tiv ity w i tho ut any s ta te
interference.
G end ered labor m arkets a re no t new , of course . Sex segregat ion and
wa ge inequal i t ies hav e been wel l -doc um ented in Poland as wel l as in o ther
cou ntries of the region (Einhorn,
1993;
Re szke , 1989). D esp ite the rhe tor ic,
i t is easy to sh ow that labor force part icipa tion did no t em an cip ate w om en
un de r s tate social ism, if by this term on e m ea ns econom ic equ ali ty b etw ee n
m en and w om en. Sta te soc ia l ism did , how eve r , c rea te a un iqu e conf igura
t ion of feminized labor m arkets . Professions requ ir ing high er educ at ion,
for example, part icularly those in the human services , tended to be femi
nized, if no t absolute ly , a t leas t dram at ica l ly m ore so th an the i r cou nter par ts
in the we st . M edical and den tal care, law, econo m ics, an d ho st of posi t io ns
in midd le - l eve l m anag em ent were dom ina ted by wo m en. W om en he ld
m any jobs in econom ics and plan ning , and in accou nt ing and adm inis t ra -
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 296/375
79
t ion w ithin s tate enterp rises . To be sur e, these posi t ion s rarely entai led the
responsibil i ty, s tatus , or the salary of equivalent job t i t les in western
com pan ies ; un de r centra l ized plann ing, such pos i t ions we re typica l ly fai rly
rou tine an d eve n clerical. M oreov er, relat ive to skil led ind ustri al jobs , the
service sector as a w ho le wa s s tarv ed for fund s H eyn s an d Bialecki , 1993;
BiaJecki and Heyns, 1993).
The t rans i t ion to a market economy in Eas tern Europe wi l l involve
dramatic changes in both labor markets and s trat if icat ion more general ly.
Whether one at tr ibutes change to the collapse of s tate social ism or to the
dawn of a market economy, the emerging pa t te rns of employment offer a
rare opportuni ty to observe the sources and consequences of gender in
equ ali ty in the labor ma rket . Edu cation al inst i tut ions of al l sorts are u nd er
cons iderable pressure to re form the curr iculum, and to educate a new,
m arket -o r iented e l ite . Yet as w e ha ve seen, educat ion al prog ram s, inc lud
ing those in bu sine ss and econo mics, tend to recruit s tude nts from the sam e
socia l gro up s as before . Current ly , w om en are as we l l - represented in h igh er
edu catio n and in econ om ics as they we re in the past . Desp ite dram atic
increases in the propor t ion of s tudents s tudying bus iness and economics ,
the percentage of women is s t i l l high, i f western patterns are taken as the
no rm . I ronica l ly , m an y recrui te rs have been no np luss ed by the sex co m po
si t ion of many jobs and professional t raining programs.
1 3
Each of the business schools s tudied enrolls roughly equal numbers of
w om en and m en, and these pro gra m s are no t a typical for Po land as a whole .
As w e ha ve seen, ho we ver , the va lues and aspi ra t ions of w om en s tud en ts
differ from those of the me n, part icu larly w ith respect to s tart ing their ow n
bu sines s . The que st ion bec om es w he the r w e can account for gender-specific
aspi ra t ions am on g s tude nts by di fferences in bac kg rou nd and va lues .
Ult imately, of course, occupational mobil i ty depends on factors not antici
pa t ed by s tud ent s and not me asu red in th is research. But w e can begin to
un rav el the sou rces of ge nd er difference in at t i tu des and v alue s tha t pred ict
ent repreneur ia l aspi ra t ions .
Table 12.4 presents the complete l is t of occupational dest inations sepa
rately by gend er for the Polish bu sine ss s tu de nts . As this table m ak es clear,
both men and women bus iness s tudents aspi re to d iverse occupat ional
roles.
The w om en are at least as l ikely to favor priv ate f irms as the me n, a nd
both m en and w om en overw helm ingly reject s ta te enterpr ise . M oreov er ,
female s tudents do not seem indecis ive about their goals or career plans;
they are at least as likely as the men to have formed specific ambitions
during their ini t ia l year, and these ambit ions resemble those of men s tu
de nts in m ost respec ts . Th e s ingle i tem that s ignificantly di s t ing uish es m en
and w om en s tuden t s is thei r p ropen s i ty to a sp i re to ow ning an d run n ing
their own businesses . Almost half of the men, compared to less than one
w om an in three hope s to be the prop r ie tor of the ir ow n f irm.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 297/375
28
TABLE 12.4 Career Asp irations and Challenges by Gender,
Polish Business Students, 1992
PERSONAL ASPIRATIONS
1
Ow n one's own business
Private Firm
State Firm
Foreign Firm or
Joint Venture
In Poland
Abroad
Research, Continue
Education
Don't know, No A nswer
Males
45.4
38.0
11.8
31.4
12.7
21.8
7.0
11.4
Females
29.9
40.2
6.7
40.2
16.0
17.5
6.2
9.3
Total
38.3*
39.0
9.5
35.5
14.2
19.9
6.6
10.4
The greatest challenge for business
in post-communist Poland is
Organizing and running
a new business 41.1
Restructuring, reforming
state enterprise 29.4
Other, Do n't Know 29.5
The primary reason for
attending business school
is to make money 63.8 50.0 57.5
'Students were allowed to choose multiple answers to the question of
occupational destinations, hence the percentages do not equal 100 .
This is the only sex difference that is significant, (p < .01).
Polish business schools offer a unique laboratory for comparing gender
differences in aspiratio ns an d values. H istor ically, bu sines s an d e cono mic s
at t rac ted mo re wom en th an me n; hence , w om en w ere avai lable as both role
m od els an d pee rs to a m uc h greater extent than is true in the we st . N eith er
minor i ty s ta tus nor token ism wi th in bus iness and m anage me nt w ou ld be
expected to dep ress Polish w om en 's aspira t ions and achieve men t , as seems
to occur in western firms (Kantor, 1977).
There are several al ternative explanations for the relat ionship between
ge nd er and ow nersh ip , and som e evidence in favor of each. O ne m ight
arg ue that wo m en are m or e realis t ic abo ut their career pla ns than me n, a nd
less l ikely to aim for dis tant , unreachable goals . Perhaps women's aspira
t ions are more dependent on their family of or igin than those of men.
Finally, sex-specific economic values could also influence the differences.
Paula England (1993) and other feminists have argued that neo-classical
economics assumes a va lue s t ruc ture tha t i s in imical to women's specia l
con tributio ns, by focussing on self-interest an d on rational, profit-max imizin g
36.6
29.3
34.1
39.0
29.3
31.7
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 298/375
281
beh avio r that ign ore s conce rns for the collectivi ty. The w om en in this
sam ple cons is tently scored lowe r than m ale bus in ess s tud ent s on laissez faire
valu es , a l th ou gh th ere were no s ignificant gen de r differences in the na tion al
sam ple . The que s t ion becom es can the d i ffe rences in pro pr ie torsh ip be
explained by a gender-specific configurat ion of values , aspirat ions, or
s tatus .
A very large l i terature on gender differences in achievement has aimed
to d is t inguish be tw ee n aspi ra t ions , va lues , an d parenta l s ta tus . (Marin i ,
1978; H ou t and M org an , 1976). E xplana t ions tend to em pha s ize d i ffe rences
in social izat ion an d aspirat io ns as the key variables . H ow ev er, there are
pers is tent d i f fe rences in a t ta inment be tween men and women tha t a re
re la tive ly in de pe nd en t of e i ther bac kgr oun d or aspi ra t ions . W hi le the
Polish business s tudents are a relat ively select group, family resources ,
va lue s , an d aspi ra t ion s predic t ow ner ship aspi ra t ions . H ow eve r , family
income and sector of employment explain differences in values at least as
we l l for m en s tud en ts as for w om en s tude nts . M oreover , bo th m en an d
women s tudents grew up and were soc ia l ized in a soc ie ty tha t d id not
end orse l ibera l econom ic va lues or bus iness aspi ra t ions ; henc e , there seems
no reason a priori to expec t tha t the re la t ionship be tween va lues and
aspirat ions would differ by sex.
The model l ing s t ra tegy adopted was based on extending the logis t ic
equ at ion prese nted in Table 12 .3 . Ho w ever , each of the ind ep en de nt
var iables wa s redef ined as a gen de r in te rac t ion . Predic t ing the log od ds of
aspi r ing to ow n o ne s ow n bus iness w as assum ed to be a func tion of soc ia l
background, pr iva te school enrol lment , the perce ived cha l lenge of propr i
e torship , the des i re to make money, and the measures of laissez faire and
collectivist va lue s. To som e exten t, each of these factors is different for m en
a n d fo r w o m e n . M a l e s t u d e n t s a r e m o re i n f lu e n c e d b y h a v i n g a
self-employed parent or by enroll ing in the private school, a l though the
direct ion of effects are s imilar . Women s tudents are somewhat less l ikely
to perce ive ow nin g a bus in ess as the grea tes t cha l lenge , an d cons id erably
less l ike ly to admit making money was the pr imary reason they enrol led .
H ow ev er, these variab les influence pro pr ieto rsh ip for bo th sexes , an d their
effects are largely addit ive . M eas ure s of valu es , ho w ev er, influence asp ira
t ions very differently for m en and for w om en . W om en are .2 of a s ta nd ard
deviat ion less laissez faire than m en, and there is l it tle relat io nsh ip be tw een
scores on th is var iable and any m easu re of aspi ra t ions . In fac t, w om en w ho
want to make money score s igni f icant ly lower on laissez faire va lues than
women who do not . Individual ism, in contras t , does not d is t inguish men
from women; however , col lec t iv is t va lues opera te to inhibi t aspi ra t ions
only amo ng w om en. The com binat ion of these tw o fac tors, ope ra t ing in
ta nd em , is sufficient to red uc e the effect of ge nd er to insignificance. A fully
specified gender model , with al l effects defined as interact ions, is not
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 299/375
8
TABLE 12.5 Logistic Regression of Ownership Aspirations on Student Back
ground and Values, Polish Business Students, 1992.
Independen t
Variables
Self-employed
Private school
N o .
Dest inat ions
Cha l l enge -Own
Money
Laissez-faire-Men
Ind iv idua l i sm-Women
Sex
Constan t
-2 Log Likelihood
Model Chi -Square
Goodness of Fi t
Coefficient
.581
.591
.390
.740
.401
.236
-.370
.354
-2.208
504.130,
58.886
421.859
(S.E.)
.260)
.316)
.107)
.222)
.223)
.194)
.179)
.284)
.307)
P
.025
.061
.000
.001
.072
.225
.039
.212
.000
Exp(B)
1.787
1.806
1.477
2.096
1.493
1.266
.691
1.425
w ith 414 degr ees of freedom .
(p=.3842)
signif icantly better than the m od el with only valu es al low ed to interact w ith
ge nd er . This m od el, given in Table 12.5, imp lies that ba ck gro un d, careers
chosen, chal lenges perce ived, des i res to ma ke mo ney , and p r iva te school
enrollment s ignif icantly inf luence ownership aspirations for both men and
women. Sex-specific values, including the positive effect of
laissez faire
o n
men and the negat ive inf luence of individual ism, account for gender
differences . A m on g these Pol ish bus iness s tudents , va lues su pp or t in g
collectivism diminish economic aspirations, but do so only for women.
Despite the fact that for both men and women, the correlat ion between
ind iv idua l i sm and
laissez faire
att i tudes is posit ive, and for both men and
women, background is equally predictive of values, their effects are quite
specific by gender.
C onc lu s ions
M ost of this pa pe r ha s been relatively descr iptiv e, with l i tt le at ten tion to
the larger issues of the econ om ic transit ion in East Cen tral Euro pe. Pe rha ps
con clud ing on a m ore specu lative note is , therefore, justif ied. O ne m igh t
well ask whether the insti tutions s tudied have much to do with the labor
m ark ets and training insti tutions in s table pos t- indu str ial societies , that are
not in t rans i t ion. O ne ques t ion, in par t icular , recurs : Are bus in ess schoo ls
ration al , as a m ea ns for creating or joining a m ark et econom y? For m ost
of this pa pe r , the intentio ns of the schools an d s t ud en ts hav e been taken at
face value. The analysis focussed on the consistency of liberalism, and on
the determinants of career decis ions.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 300/375
83
From an economic point of view, post-comm unist bu siness schools seem
rather paradoxica l. The economic failures of socialism are seen in the lack
of market rationality linking supply and d em and. W hether one sees
plann ing as too complex or the state as too pow erful, the failure of socialism
is the same. M arkets, in turn , rest on indiv idua l effort and com petition,
without, if possible, state intervention. Entrepreneurship and innovation,
two systemic deficiencies of socialism, involve skills and risk-taking that
cannot be taught. To create a mark et economy, one needs private prop erty
and a relatively stable polity to protect individual rights; incentives for
self-interested, profit-maximization are assum ed to be innate. Credentials,
in contrast, create possibilities for monopoly, closure, and exclusivity.
Establishing schools for training entrepreneu rs and bu siness leaders w ould
seem to be icing the cake before it is baked.
At the outset, I argued that ma rket transition really means a period of
time or a place of social change, involving the supposition that markets
should replace state control in every sphere; as such, they provide a
perspective on how societies and ideologies are organized and how they
change. On the one han d, there is the universal assumption that market
processes will rationalize labor markets and make them m ore competitive
and efficient. On the other, when the process
is s
extreme
s in
East Central
Eu rope, established ins titutional processes have all but collapsed and n ew
incentives and m echanism s for job placement have yet to emerge. Educa
tion, in particu lar, has been divested of a major m otivational incentive, the
power to determine the economic fate of their stud ents . Althoug h indiv idu
als are now free to save or spen d their resources, and to plan their futures
without state interference, the information that would make such calcula
tions possible is no t yet available. Ironically, how ever, education, and
particularly bu siness education, is expanding throug hout the region.
From an institutional point of view, post-communist business and eco
nomics education is grow th industry . The new or reformed schools claim
to have the expertise needed for new era; moreover, the dem and for their
services has never been greater. Their traditional mission, how ever , their
faculty, and v irtually all of their assets are inherited from
the past
The basis
of their power and status has, however, undou btedly shifted. The schools
have lost a privileged relationship to state planning as well as their mo
nopoly of selection and placem ent. W hat they retain, and which apparently
is their most critical resource, is the claim to possess and transm it special
ized expertise in economics. The schools claim to be sure, that they have
reformed to meet the new challenge, yet their resource base is the same as
before. They reform not by hiring new faculty or discharging the old,
building new facilities or developing new curricula; this would be costly.
Instead , they liberate their resources from the state, m arketize their
prod ucts, and transform their administrative structures. The most visible
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 301/375
84
aspects of reform consis t of rhetorical debates and untested claims of
bus iness perspicac i ty and competence .
From the poin t of v iew of s tud ents , the sys tem w orks in m uc h the sam e
w ay a s before, but wit h a ver y different rhetoric . The re are, to be su re, n ew
and diverse opt ions and opportuni t ies for t ra in ing, a l though these are
uniformly qui te expens ive . I t would be unders tandable i f s tudents were
ra ther dubious about the c la ims or the miss ion of economics educat ion.
Pr olo ng ed training req uires com m itm ent, an d in this case, i t is n ot clear
w he re i t wil l lead. The credibil ity of econo m ics could be cha lleng ed, as w ell
as the autho ri tat iv e claims of their ins truc tors , especial ly w h en t he retu rn s
to schoo ling are s ti ll unclear. The incentive to invest in h u m a n capital
presumes tha t one can ca lcula te an expected re turn , d iscount ing the los t
op por tun i t ies of ear ly m arket ent ry . The pr im ary m ot iva t io n of the
bu s ines s s tude nts , as w e ha ve seen, i s to m ak e mon ey. Yet ins tead of
ac tua l ly s t a r t ing a bus ines s o r s eek ing employment in the expand ing
private sector, they choose to s tudy the subject , spending scarce resources
in the ho pe of acq uiring k no w led ge a nd creden tials that wil l , a t least in the
long run ,
pay-off.
The ra t ional i ty of e i ther providing or inves t ing in bus iness educat ion
rests on the convict ion that with knowledge and expert ise , especial ly if
autho r i ta t ive and accredi ted , one can un de rs ta nd , m anip ula te , and control
m ark et forces. The idea tha t educa t ion and informed p lann ing can ou tw i t
the market was common in virtually al l social is t countries , part icularly
thos e that ceased to rely on terror or ove rt coercion. M oreo ver, e du ca tion
was assumed to be the best s trategy for reforming the system, and for
crea t ing ne w values and new socia li st c i t izens . Pos t -co m m unis t bus iness
schools have not changed this conception of their role; the content of the
message ha s changed , bu t the a s sum pt ions abo u t pu rpo se an d mis s ion a re
m uc h the same as before . W hy does th is idea pers is t , w he n so m an y other
aspe cts of socialis t ideo logy ha ve been forcefully discred ited?
The answ er , I be l ieve , is cent ra l to un de rs ta nd ing the m ark et t rans i t ion .
Faith in the magical power of schooling to create new personali t ies and a
new socia l order i s not unique to pos t -communis t countr ies ; educat ion is
heralded as the panacea for every kind of reform, especial ly in t imes of
na t ion al c r is is. W hat is un iqu e is the eno rm ou s econom ic un cer ta in ty
prevail ing in East Central Europe and the claim that special ized business
exper t i se i s the mos t appropr ia te cure .
For bo th bus in ess s tud ents a nd bus ine ss schools , concep t ions of exp er
t ise serve to redu ce unc ertain ty. The re is , to be sur e, the ad de d incentive of
possibly being the va ng ua rd of a ne w , s ti ll dim ly envis io ned , social o rde r.
For the faculty, few al tern atives bu t reform are possible . For the s t ud en ts ,
how eve r , choos ing to a t tend bus in ess school involves a ca lcula ted gam ble .
Education no longer offers a clearly planned career trajectory; the estab-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 302/375
285
l ished l inks between the skil ls learned and the jobs assigned have been
term ina ted . But there is the possibili ty, ho w ev er s l ight , of vast r iches just
aro un d the be nd . W hile the schools no longer offer cer tainty ab ou t career
prospects and rewards, they can and do furnish examples of successful
entrepreneurs , and they c i rcula te adver t isements for pos i t ions wi th unbe
lievably hig h salar ies . H ow ev er remo te these possibil i t ies m ay be in
actuali ty, an d how ev er ir rat ional an d chaotic the curre nt m arke tpla ce in fact
is , business schools reduce the r isk of unemployment at least in the short
ru n, and the y offer the possibil i ty of im m en se we alth in the future. Business
edu cat ion is not ra t iona l , if by th is we m ean the expected re tu rns are
greater than the costs that are l ikely to be incurred, ei ther socially or
ind ivid ua lly. Th ese societies r isk creating far m ore dis i l lusio ned gr ad ua tes
tha n successful en trep ren eu rs . At the sam e t ime , their pr im ary function
m ay be to redu ce the uncer ta int ies inherent in ma rket processes , ra ther than
prepar ing en t r epreneur s .
otes
Fun ding for this research was provided by the National Council on Soviet and
East European Research. A preliminary report of this research was presented at the
M idterm conference of the Sociology of Education section in Am sterdam , July 1992.
1.
Between 199 and
1992,
the num ber of schools of higher economics education
increased from five to fourteen, and enrollments from 24 to 4 thousand in Poland.
By 1993, 14.6 percent of all post-secondary students in Poland were studying
Economics, com pared to 10.6 percent in 1990-91.
Szkaty Wyzsze w Roku Szkolnym
1992/93
(Warsaw:
Glowny Urzad Statystyczny
1993).
2.
In recent decades, the dea rth of training m ay have been a practical response to
labor marke t plann ing , as much or more than ideological distaste. Socialist coun
tries did provide specialized training in business and economics when needed.
Agricultural universities, for example, offered courses in economics and ma rketing ,
because farming in Poland remained overwhelm ingly private. Dep artments of
foreign trade provided instruction in western banking, trade regulations, and
international finance sufficient to handle im ports and exports.
3. A fourth site for business training m ight be the Polytechnic University. In
Warsaw, the
Wydzial Mechaniczno technologiczny
trained m echanical engineers; this
school recently announced their intention to admit and train m anagers for heavy
industry.
4.
Prywatna
was the second private school in Poland, the first being the Catholic
University in Lublin, which had survived the entire communist era. In 1992-93, the
Private Higher School of B anking and Insurance was approved , as well as a second
private business school in Warsaw. By 1994, thirty-three private post-secondary
program s were in operation.
5. Permission to interview studen ts was granted rather late in the academic year,
and it was not possible to sample classrooms or to follow-up absent students.
Questionnaires were distributed and collected in large introductory classes at the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 303/375
286
Management School and SGH. In the case of the priva te business school , s tu de nts
we re app roached d u r ing requ i red comp u te r workshops . The cod ing of ques t ion
na i res was d one by the Centrum Marketingu iAnaliz Spolecznych under the direct ion
of Joann a Konieczna. Q uest ion naires w ere adm inistere d to 122 stu de nts at the
Warsaw Universi ty Management Department , out of a total enrol lment of 190
first -year students. At SG H two hundred and forty students out of 720 fi rst -year
s tude nts comple ted ques t ionnai res , and
61
ou t of
220
first-year stud en ts at Prywatna.
6. As a rough comparison, a nat ional sample of universi ty students in 1991
[N=2,428] rep orted that
41
percen t had fathers w ith som e higher edu cat io n, and 20
percent worked in the private sector (Wnuk-Lipinska, 1992). Warsaw students are
not , of course , repre sentat ive of the countr y as a w hole , and private sector e m ploy
ment has grown dramatical ly since 1991.
7. The i tems w ere selected from the sho rt scales dev elope d by researc hers at
Nuffield College, Ox ford, for the British Elections Project. Th e scales w er e con
structed to m eas ure fun dam enta l dim ensio ns of pol i tical values , or stable core
beliefs. See A. F. H ea th (1991) for a des criptio n of the items and a n asses sm en t of
their reliabili ty and validity in Britain.
8 . Several exp loratory factor analyses w ith LISREL-7 w ere perfo rm ed on the
complete set of i tems. The fi rst two factors for both the nat ional sample and the
bus ines s stu den ts hav e similar st ructu res, but they are not ident ical . Form ally,
testing the equivalence of factors requires five independent tests: 1) Equality of
covar iance mat r ices ,
2
The meas urem ent m odel o r fac tor pa t te rn
is
comm on to bo th
groups , 3 ) Assuming common fac tors , equ iva lence of corre la t ions , 4 ) Assuming
equal correlat ions, common error matrices, and 5) Assuming al l of the above, the
equivalence of Phi , the variance-covariance matrix for exogenous factors (Bollen,
1989). For the
laissez faire
scale, th e first hyp oth esis is rejected, the sec on d a nd thir d
are accepted , wh ile the final tw o tests mus t be rejected. He nce, one conclud es th at
a l though the sca les a re h igh ly corre la ted , the two samples have d i f fe ren t
variance-covariance matrices and different error st ructures. Since ident ical scales
we re need ed in order to com pare the de term inants of va lues be tw een sam ples , the
i tems w ere po oled, and a single factor extracted for each of the three dim ens ions . For
both samples, the pooled factor scores correlated with the principal components
very highly, and never less than .935. The means and variances for each scale ,
how eve r, are specific to the pooled sam ple, and hence the stu de nt scores are g reater
than 0 and the variance less than 1.
9. Th e items for the Laissez faire-Socialist scale w er e 1) O rd in ar y w or ki ng
peo ple do no t ge t
a
fair sha re;
2
There
is
one law for the rich and o ne for the po or ;
3) M ana gem ent-la bor coop erat ion is imp ossible; 4) Big busin ess benefi ts own
ers ; 5) Go vernm ent should equal ize incomes; 6 ) M anag em ent explo i ts l abor ;
and 7) Pe ople are po or bec aus e of social injustice.
10.
The items included in the Libertarian -Au thori tarian Scale w er e 1) Schools
should teach obedience ; 2 ) The pun ishm ent for b reak ing the law sho uld be m ore
seve re; 3) Y oung peo ple toda y lack respect , 4) For som e crime s, the de ath
pen al ty is ap pro pri ate; 5) The law m us t be obey ed; and 6) Ce nso rship is
necessary to uphold mora l s tandards .
11 . The i tems m eas urin g Collect ivism -Individual ism w ere 1) W elfare benefi ts
are too gen ero us; 2) Pov erty is usual ly a pe rso n 's ow n fault ; 3) M any peop le
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 304/375
287
who get social benefits do not really deserve them; 4) When someone is unem
ployed, it is usually their own fault; and 5) Governm ent gives too many hand
outs.
12.
Other responses to this item w ere that they had been interested in business for
some time (8%), that this particular program had been discussed (8%), or that it had
been recommended to them (7%), that their parents (2%) or employer (1%) had
suggested it. Twelve percent of the studen ts chose another reason .
13. As part of this project, a number of western business executives and represen
tatives of internationa l foundations w ere interviewed. Many expressed concern at
the large num ber of wome n in economics and management program s; affirmative
action for men m ight be necessary, we were told. The pattern w as interpreted by
our informants as indicative of the low status of business and commerce under
communism; enhancing the status of these training programs might require dis
crimina ting in favor of men.
14. The faith in education was wide spread, but p erhaps not as pervasive as the
faith in free markets. At least one faculty m ember com mented that educational
reform w as Utopian, and hearkened back to end uring tenets of socialism. His
comm ent was that Poland needed capital, not capitalists.
eferences
Adam ski, Wladysiaw W., Ed. 1993. Societal Conflict and Systemic
Cliange
Warsaw:
Polish Academy of Science, Institute of Philosophy and Sociology.
Beksiak, Janusz , Ewa Chm ielecka, and Urszula Grzelonska. 1991. Kierunkizmian W
wyzszym szkolnictwie
ekonomicznym
[Direction of Changes in Higher Economics
Schooling], Warsaw: M aterialy Fundacji im Stefan Batorego.
Biatecki, Ireneusz and Bogdan Mach. 1992. Social and Economic Orientations of
Polish Legislators Against a Background of the Views of Polish Society,
The
Polish Sociological Bulletin 2:167-186.
Bollen, Kenneth A. 1989. Structural Equations
with
Latent
Variables.
New York:
Wiley.
Brubaker, Rogers. 1984. The Limits
of Rationality: An
Essay on the Social and Moral
Thought
of Max
Weber. London: George Allen and Unwin.
Chmielecka, Ewa. 1993.
Katalog Szkoiy Glownej
Handloivej, [The SGH C atalogue]
Nauka i
Szkolnictwo
Wyzsze
1. Warsaw: Centrum Badan Polityki Naukowej i
Szkolnictwa Wyzszego.
. 1993. M arket Economy and
the
Reform of Education
in the
W arsaw School
of
Economics,
Pape r presented
at the
Conference, Change in
the
Systems
of Higher
Education
in
Central European
Countries
Madralin, Poland, October.
Collins, Randall. 1980, Weber's Last Theory of Capitalism: A Systematization ,
American Journal
of
Sociology 5 (December): 925-942.
Converse, Philip E. 1964. The Nature of Belief Systems in Mass Publics, Pp .
206-261 In Ideology and Discontent, edited by David E. Ap ter. New York: Free
Press.
Einhorn, Barbara. 1993. Cinderella Goes to
Market:
Citizenship
Gender,
and Women's
Movements
in
East Central
Europe.
New York and L ondon: Verso Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 305/375
288
Eng land, Paula. 1989. A Fem inist Cri t iqu e of Rat ional-Choice The ories: Implica
tions for Sociology, American Sociologist 20:14-28.
En glan d, Pau la. 1993. The Sepa rat ive Self Androcentric Bias in Neoclassical
A s s u m p t i o n s , I n eyond
Economic Man: F eminist Theory and Economics,
ed i ted by
M arian ne A. Ferber and Jul ie A. Nelso n. Chicag o, Un iversi ty of Ch icago P ress.
Edi tors . 1994. The Flawed Educat ion of the Euro pean Businessm an,
The Econo
mist 331, 7866 (June 4): 61-62.
Grayson, Leslie E. 1993.
Paving the
Road
to Prosperity: Man agement Education in the
Czech Republic, Slovakia, Hungary, and Poland Laxenburg , Au st r ia : In te rna t iona l
Inst i tute for Applied Systems Analysis , August .
H aus er, Ewa, Barbara H eyn s, and Jane M ansb ridge . 1993. Fem inism in the
Interst ices of Pol i tics and C ulture: Poland in Tran si t ion, In
Gender Politics and
Post-Communism,
ed i ted by Na net te Funk and Ma gda Muel le r . N ew York:
Rutledge Press.
Hea th , An thony F. 1991. The M easu rem ent of Co re Beliefs and Va lues, W ork ing
Pa per #2, Nuffield College, Oxford (Au gust) .
H ey ns , Barbara and Alek sand ra Jasinska-Kania. 1993. Values, Pol i t ics, and the
Ideo logies of Reform: Pola nd in Tra nsitio n, P p. 169-194 in Democratization in
Eastern
and W estern Europe: A
Research
Annual,
Vol 1, ed ited by Fre derick D . W eil,
(Greenwich, Connect icut : JAI Press.
H eyn s, Barbara and Jan Szczucki . 1992. Ed ucat in g En trepr ene urs: V alues, Edu ca
t ional Reform, and the N ew Polish Business Schools. Pap er pre sen ted to the
Interna t ional S ociological Associat ion, Research Com m ittee on th e Sociology of
Educat ion , Midye ar Conference , Ju ly, Am sterda m , Ne ther lands .
H ou t , Michael an d W. R. M org an. 1976. Race and Sex Va riat ions in the Ca use s of
the Expected Atta inm ents of H igh Schools Seniors,
American
Journal
o f Sociology
81:364-94.
Ing leha rt, Ro nald . 1988. The Re naissa nce of Political C ult ur e: Basic V alue s,
Pol i t ical Economy and Stable Democracy. American Political Science Review
(December) .
Kan tor , Rosabeth . Men and W omen of the Corporation. N ew York: Basic Books, 1977.
Kov acs, Janos M., Ed. 1991.
Red iscovery of
Liberalism
in Eastern Europe,
Special Issu e
of
East Euro pean Politics and
Societies
5,
1.
M arini , M argar et M. 1978. Sex Differences in the D eterm inan ts of Ad olesce nt
Aspirat ions: A Review of Research, Sex Roles 4: 723-53.
Mii l ler, Ale ksan der. 1991. Ch ang es in the System of Hig her Econom ic Educ at ion
in Poland : The Case Stu dy of Szkoia Gtdwna
Handlowa,
in
Transformation ofScience
in Poland
edi ted b y Anton i Kuklinski , (Warsaw : State Co m m ittee for Scienti fic
Research, 1991).
Pa ul, Ellen Franke l, Fred D . Miller, Jr., an d Jeffrey Pau l, Eds. 1993
Liberalism
and the
Economic Order. N ew York: Ca m bridg e Univers ity Press .
Reszk e, Irena. 1989. Sex Inequal i t ies in the Labo ur M arket in Po land ,
Sisyphus
6:
83-102.
Ro keac h, M ilton. 1979. Understanding Human Values. N ew Y ork: Th e Free Press.
Rueschemeyer, Mari lyn, Ed. 1994
Wom en in tlw Politics of Postcommunist Eastern
Europe.
A rm on k, N ew York: M. E. Shar pe, 1994.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 306/375
289
Szacki, Jerzy. 1990. A Revival of Liberalism in Po land?
Social Research
57, 2
(Summer) :463-491.
Szelenyi, Ivan
in
collaboration w ith Robert Manchin, Pal Juhasz, Balint Magyar, and
Bill M artin. 1988.
Socialist Entrepreneurs: Embourgeoisement in Rural Hungary.
Madison, Wisconsin: University of Madison Press.
Weber, Max. [1904-1905] 1930. The
Protestant Ethic
and the Spirit of
Capitalism
[Translated by Talcott Parsons]. New York: Scribner's.
Weil, F rederick. 1993. The Development of Democratic Attitudes in Eastern a nd
Western German y in a Com parative Perspective, Pp. 195 225 in
Democratization
in
Eastern and Western Europe:
A
Research
Annual Vol 1, edited by F rederick D.
Weil, Greenw ich, Connecticut: JAI Press.
Wnuk-Lipiriska, Elzbieta. 1992. University Student Survey ,
ulletin of the Centre
for Science Policy and Higher Education
University of Warsaw.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 307/375
This page intentionally left blank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 308/375
PART FOUR
Social System ontexts
There is no sh arp distinction be tween the pape rs in Part IV and at least
two of those in Pa rt III. Both DiPrete and M cManus and Hey ns situate their
analyses and interpretations in the na ture of societal system s. H ow ever , the
pap ers in Part IV have a somew hat broader concern with the relevance for
mobility patte rns of the varied and ch anging natu re of the societal system s
within which stratification processes occur.
In the first paper in Part IV, Michael Hout points to inter-societal
variations
in
social mobility and suggests ways
in
which actions by the state
can affect the degree and kind of mobility observed. He suggests one
particular way in which actions of the state can indirectly alter mobility
patterns, namely, through the level of sup port of post-secondary education.
He describes the different responses of three nations (Ireland, Italy and
Russia) to the post-World War II increased demand for university educa
tion an d ho w the three appro aches affected the degree of access different
social classes had to higher education .
e
argu es that past actions of the U.S.
state and nationa l governments have increased equal access and facilitated
intergen erationa l mobility. But he sees current policy as threa tening to
redu ce access for potential stu den ts in the lower classes. Na tional higher
education policy is thus seen as a very significant factor in de term ining the
degree of op en ness of a society's mobility pattern.
Seym our S pilerman and Hiroshi Ishida use their study of a large Japa
nese financial firm to point up several ways in which a society's cultural
norm s affect the career patte rns of the firm's w orke rs. There are sharp
contrasts between this Japanese firm and com parable A merican firms in the
role of gender in occupational placement, the firm's recruitment methods,
the firm's commitment to workers, the degree of specialization, and the
structu re of rew ard s. Given these differences, Spilerman and Ish ida's
primary concern is to understan d promo tion patterns w ithin the Japanese
firm's m anagerial work force. They find a very grad ual sifting process
rather than a tournament or cumulative advantage patterning of careers.
Strong Japanese cultural norms of social harmony, respect for elders, and
putting collective goals before personal ones help explain both the struc-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 309/375
292
rural features of the firm, and the career experiences of workers.
Also making comparisons between societies, Arne Kalleberg's paper
sugg ests ways in which curren t and recent shifts in the wo rld econom y can
affect the way firms op erate in different societies. He argues tha t firm
flexibility is increasingly necessary for successful world com petition .
Using various societal contrasts, he show s how a society's institutional and
political-economic characteristics affect the ease with which various kinds
of flexibility can be introdu ced . Assum ing a grow ing need for flexibility,
Kalleberg develops two sets of possibilities for the future: an optimistic
scenario foresees an overall up -grading of the skill level of the labor force,
bu t a pessimistic scena rio foresees increased polariza tion of the labor
force into
small high skill flexible core and a large residua lized per iphery .
He specu lates abou t the conditions that might lead to one or the other.
The papers in Part IV identify societal and inter-societal characteristics
that affect the distribution of the popula tion into stratified
levels.
Spilerman
and Ishida show ho w a society's cultural norm s and definitions prov ide a
matrix of forces that help
to
define possible and app rop riate moves in the
labor force. Hout demonstra tes how national political decisions can enlarge
or limit stu dent s' access to high level educational credentials. And, finally,
Kalleberg discusses ways in which inter-societal economic relationships
can feed back on intra-societal institutional arrangements and the labor
force experiences of a society's worke rs. All of these pap ers show how
societies shape the flow of individuals into their stratification systems.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 310/375
3
The Politics of Mobility
Michael Hout
University
of
California erkeley
Social mobility gauges the openness or closure of a society. An open
society will exhibit a high level of social mobility d ue to a low persistence
of privilege from one generation to the next; a closed society will exhibit a
low level of social mobility due to persistence of social position across
gene rations. As citizens and social scientists we endo rse equality of opp or
tunity even wh en w e disagree about the appropriate absolute level of
inequality so openness is a widely shared goal. We might even say that
although w e stud y mobility for many reasons none is more imp ortant than
to study social mobility in order to change it for the better. That said we
might
go
on
to
ask how the accumulation of research on social mobility over
the pas t decade guides that interest in mak ing society better.
Research over the past decade has contradicted several widely cited
conclusions from earlier research and speculation. Among the key new
findings:
>Social mobility varies significantly among societies.
>Social mobility has changed over time in most societies—although
change over time is less than differences am ong societies.
>Variability in social mobility—once thou gh t to e due more to so-called
structural mobility—in fact reflects differences in openness.
>Variability in openness— once tho ugh t to reflect certain law -like prop
erties of industrialization—in fact reflects political choices.
>Two kin ds of political choices can influence the degree of open ness in
society. Capitalist social welfare policies can pro m ote openness by pro vid
ing a w ide range of social service
jo s
while mitiga ting the effects of losing
out in the jo s com petition by redistributing incomes. Free marke t em ploy
m ent and investm ent policies typically increase income inequality but they
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 311/375
94
can nonetheless foster openness if the s tate pursues aggressive investment
in higher education (which in turn directly fosters openness) .
I point only to the lessons of the past decade's research for political
sociology in this chapter , so my reading of the research l i terature is re
s tr icted. Breiger (1995) provides a more comprehensive review.
Stru ctu ral m obility is the outc om e of factors that affect all orig in c lasses
pro po rtion ally . For the m ost par t , s tructu ral mo bili ty occurs bec aus e som e
classes gr ow w hile othe rs shri nk ov er time, e.g., the historica lly significant
decl ine of agr icul ture and the gro w th of profess ional man age r ia l e m ploy
m en t has increased the op po r tun i ty for up pe r whi te col lar em plo ym en t for
all workers regardless of their origin class.
ef in it ions
Social mo bili ty occurs betw een social classes . M uc h of the researc h ov er
the last decade has addressed the issue of def ining classes (Wright 1985;
Erikson a nd G old tho rpe 1992a). H ow ev er , if w e s tep back from this (ad m it
tedl y im po rtan t) detail we can dr aw conclusion s that apply to the w or k of
poli t ical sociologists so long as any reason able class scheme is em plo ye d. By
r e a s ona b l e I m ea n that the schem e contain s suff icient inform ation abo ut
the hierarchy of posit ions in society to be useful while at the same t ime
making concess ions to the amount and kind of da ta avai lable and other
exigencies of research practice. Of course, it is possible to have serious
dis ag ree m en ts on these m atters (e .g., W right 1987; H ou t and H au se r 1992;
Erikson and G old tho rpe 1992b). I wil l d o m y best to focus instead on the
points tha t the major neo-Marxis t and neo-Weber ian approaches have in
c o m m o n .
1
Mobili ty s tudies commonly feature the concepts of s tructural and ex
change mobili ty. Structural mobili ty ar ises because the marginal dis tr ibu
tions of origin s an d de stin atio ns typically differ, m ak in g it im po ssib le for all
persons to have a destination in the class of their origin.
2
Typically,
s tru ctu ral mob ili ty refers to al l those factors that are ind ep en de nt of or igins
bu t tha t also account for al l differences b etw een the dis tr ibutio ns of or igins
an d des t in a t ions (Featherman a nd H auser 1978; Sobel , Ho ut , and Du ncan
1985). By con ven tion, exchan ge mo bili ty is that par t of the mo bili ty proce ss
that produces equal f lows between or igin-destination pairs (Sobel et al .
1985).
Structural mobili ty plus exchange mobili ty does not necessar i ly
equal total mobili ty. Some asymmetry in the association between or igins
and des t ina t ions can prod uce ad di t ional mobi li ty .
3
The Sobel et al. frame
work for s tudying s t ruc tura l and exchange mobi l i ty depar ts f rom pr ior
ap pro ach es in three beneficial
w a y s :
1) it avo ids the tende ncy to view either
structural or exchange mobili ty as the residual lef t over when the other is
accounted for , (2) i t provides a one-to-one correspondence between con-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 312/375
95
cepts and measures, and (3) i t reor ients thinking about s tructural mobili ty
by replac ing a socie ty-wide concept wi th one tha t appl ies to each occup a
tional catego ry. Recent wo rk has tend ed to dr op the s trict SH D form ulation
of the problem and to adopt the def init ion of s tructural mobili ty that
stresses that it is the sum of only the factors responsible for dissimilarity
between the marginal d is t r ibut ions of or igins and des t ina t ions tha t a re
independent of or igins .
Exchange mobili ty is the symmetr ical par t of the association between
origins an d d estin atio ns (Sobel et al . 1985; Ha use r a nd Grusky 1988; H ou t
and Hauser 1992) . The
ij
terms capture al l of that association, but do so
relatively inefficiently. A w eal th of m od el s offer m an y m or e efficient w ay s
of express ing the associa t ion between or igins and des t ina t ions (e .g . ,
G oo d m an 1972, 1979, 1984, 1991; H au ser 1979; Clo gg 1982; Log an 1983;
H o ut 1983,1984 ; H o u t an d H au ser 1992; Ya m agu chi 1987; Xie 1992; Breen
1993). The goal in m od el ing is
to:
(1) m ak e statistical results m o re useful for
theory and prac t ice by giving mathemat ica l express ion to a theory about
mobility, (2) improving statistical efficiency and parsimony, and (3) facili
tat ing comparisons by reducing the number of coeff icients that must be
c om pa r e d .
Since the 1960s, imm obili ty h as received special at ten tion (G oo dm an
1965).
Nearly al l data evince immobili ty in excess of that expected on the
basis of asso ciatio n in the off-diagonal
cells.
Y am ag uc hi (1983) refers to this
dis t inction be tw een immo bili ty and off-diagonal association as the sp e
cific an d ge ne ral effects of or igin on destin ation . In this pa pe r I mo dify
his term inolo gy sl ightly an d refer to dia go na l an d ge ne ral effects .
Va riation in Structural ob ility
A commonplace of the 1970s was that s tructural mobili ty accounts for
most of the mobili ty differences among countr ies and over t ime (e.g. ,
H au ser , D ickins on, Travis , an d Kof fel 1975a, 1975b; Fe ath erm an ,
Jones ,
and
H au ser 1975) . Recent research calls that conclusion in to do ub t (W ong 1990,
1992; Ho u t a nd Ha us e r 1992). Cross-nat ional s tudies tha t re ly on par am eter
es t imates—usual ly drawing on the SHD formula t ion—ins tead of decom
po sitio ns of ch i-sq ua re sho w th at class differences in the force of st ru ctu ral
mo bili ty are very large but natio nal differences are small . Figu re 13 1 (from
Hout and Hauser 1992) i l lustrates the point . For the seven Western Euro
pean and two Eas tern European nat ions tha t compr ise the core of the
CASMIN project , there is surpr is ingly l i t t le cross-national var iat ion in
struc tura l m obili ty. Figure 13.1 sh ow s the resu lts for the ni ne Eu rop ean
nat ion s in the CA SM IN projec t (Hout and H ause r
1992).
In every n ation , the
professional classes (and to a lesser extent routine white collar workers ,
technicians, and foremen) grew at the expense of the agr icultural classes .
4
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 313/375
96
FIGURE 13.1 Structural Mobility Coefficients for Nine European Nations
CASMIN project) by Class, c. 1972
Professional
or manager I
Professional
or manager II
Clerical or
sales worker
Service
worker
Employer
Petty
bourgeois
Technical
worker
Skilled
worker
Unskilled
worker
Farm
employer
Farmer
Farm
worker
r
r
• Hungary
D Sweden
• Ireland
a Poland
• Northe rn Ireland
d r i t a i n
• Germany
Q Scotland
• France
2
1
source: Hout and Hauser 1992.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 314/375
297
FIGURE 13.2 Structural Mobility Coefficients for Six Countries
by Occupational Category, c. 1972
Professional
Manager
Lower nonmanual
Skilled manual
Semiskilled
manual
Unskilled m anual
Farmer
r.
• H
z
1 —
wmmmmmm
• • 3
Sak,
t
5a
.
1 1 1 1 1 1
3 - 2 - 1 0 1 2 :
• Hungary
O P oland
• Britain
D U S A
• Brazil
a Japan
source: Wong 1992.
Scotland and Britain have less s tructural mobili ty than the rest of the
countr ies an d H un ga ry has mo re , bu t the overa l l imp ress ion is of m ore
variat io n from class- to-class tha n from nation -to-nation . A dd in g the U nit ed
States , Japan, and Brazil to the analysis (Figure 13.2) introduces more
nat ion- to-nat ion var ia t ion (W ong
1992).
The U profile resem bles th at of the
CA SM IN cou ntr ie s , bu t the level (W ong uses the 1973 US data) is s lightly
higher than tha t in Western Europe.
5
Japan h as a substan tial ly different
structural mobili ty prof i le than the other nations have (also Ishida 1993) .
The professionals grew li t t le; the managers grew almost not at al l . On the
other hand, routine white collar and skil led blue collar jobs grew more
relative to their growth (or s tagnation) elsewhere. Brazil conforms to the
CA SM IN profi le surpr is ingly wel l . Brazil lags beh ind the CASM IN coun
tr ies in the creation of man ag em en t po sit ions, and i t is s t il l in the ph ase of
blu e collar gro w th ap pro pr ia te to i ts level of econom ic dev elop m ent .
W om en s s truc tura l mob i l ity subs tant ia l ly exceeds m en s in m ost co un
tr ies that ha ve been s tu di ed (Roos 1986; H o u t 1988; Jonsson an d M ills 1993;
W on g and H au se r 1992). Figure 13.3 (from H ou t 1988) sh ow s the results
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 315/375
298
FIGURE 13.3 Struc tural Mobility Coefficients for Men and W omen in the
United States by Occupational Category and Year 1972-75 and 1982-85
Professional
self
Professional
salaried
Manager
Proprietor
Clerical
worker
Sales
retail
Craft worker
manuf.
Craft worker
constr.
Craft worker
other
Operative
nonmanuf.
Operative
manuf.
Laborer
nonfarm
Farmer
Laborer
farm
I
-
1
—
1
-
= X L p
= T
I
V
1
• Men 1972 75
D Men 1982 85
• Women 1972 75
D Women 1982 85
-5.0
-2.5 2.5 5.0
source: Hout 1988.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 316/375
299
from the United States at two time points by way of example. Because
women are underrepresented in fanning and crafts, there is substantial
structural mobility out of these classes for women. The huge structural
mobility of women into routine white collar jobs offsets the manual and
farming deficit. Structura l mobility has declined in the U.S..
This evidence makes clear that gender and class differences in s tructura l
mobility far outweigh the cross-national and temporal differences. Al
though the differences am ong countries and within the USA over time are
statistically significant, only Japan has a distinct pattern among the ten
countries I consider here - a group of countries that includes two socialist
countries and one developing n ation.
Cross National and Temporal Differences in Association
The Lipset-Zetterberg hypo thesis of
no
cross-national variation in mobil
ity (Lipset and Bendix 1959) was supplanted in the 1970s by the FJH
hyp othesis of no cross-national variation in the association betw een origins
and destinations (Featherman,
Jones,
and Hauser
1975;
Grusky and Hauser
1984; Hauser and Grusky 1988). Early returns comparing Britain with
France (Erikson, Goldthorpe , and P ortacarero 1979) and the United States
Kerckhoff
Cam pbell, and Laird 1985; Erikson and Goldtho rpe 1985)
supported FJH. Sweden and other social democratic countries stood out,
though, as having a more open mobility pattern than o ther regions (Erikson
et al.
1979;
Grusky and Hauser
1984;
Pontinen
1984;
Erikson and Go ldthorpe
1987).
The Ne therland s resembles Sweden in openness (Ganzeboom 1984;
Ganzeboom et al. 1989). The CASMIN project (Erikson and Goldthorpe
1992b has m ade cross-national differences in association abundan tly clear.
Germ any, Ireland, Hung ary, and Poland each has a distinct national variant
on the core patte rn (see Sorensen
1992).
Japan and Brazil differ from the
other cases as well (Wong 1990,1992). Furthe rmore, changes over time in
the United States (Hout 1988), Hungary (Wong and Hauser 1992), and the
Netherlands (Ganzeboom and de Graaf
1984;
Ganzeboom et
al.
1989 refute
the FJH hypothesis as it pertains to change w ithin countries.
Thus,
no t only are mobility differences du e to structural mobility sm aller
than expected, bu t mobility differences due to association are also larger
than expected.
Both diagonal and general effects differ among nations and over time.
Figure 13.4 shows the results for cross-national variation in general associa
tion from the Western European nations in the CASMIN project (from Hout
and Hauser 1992). Germany, France, and Scotland have the strongest
general association; Sweden and N orthern Ireland have the weakest. Only
Northern Ireland is a surprise in this grou p. That is where cross-national
differences in diagona l association becom e importan t. Diagonal association
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 317/375
FIGURE 13.4 Logits for Low er Professional vers us Unskilled Ma nua l Dest
as expected und er a Linear show n by lines) and Co mp osi te show n by dots) M
Cou ntry Differences in Intercepts Rem oved, Seven Western E urop ean N ations
Ireland
CWoOCTi
« 1
3.5
3.0
Z5
2.0
1.5
1.0-
.5 -
0 •
Northern Ireland
>
/ /
yS
j f
yT
-JZ
1 1 1 1 1
5 25 35
55 85 75
source: Hou t and Hauser 1992
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 318/375
302
FIGURE 13.5 M easures of General Association and Diagonal Association
for Seven Western European Countries CASMIN project)
normed to Zero Mean and U nit Variance
2.0
1.6
1 1.0
.5
.0 I-
.5
£
0)
c
S
O -1.0
-1.5
-2.0
-2.0
France
• • Sctjtland
• Britain
Sweden
Germany
Ireland
N. Ireland
-1.0 .0 1.0 2.0
Diagonal association (z-score)
source: Hout and Hauser, 1992.
in N or th ern Ire land is s ignificantly s tro ng er than in an y other na tion in the
CA SM IN set , al th ou gh the Republic of I reland is close. Figu re 13.5 sh ow s
cross-national differences in general and diagonal association in two di
mensions. The two Ir ish nations s tand out for having a combination of low
general and high diagonal association. Sweden is the only nation in the
CA SM IN set wi th low gen era l and low diag onal associa t ion. No tably , no ne
of the CASMIN nat ions has high genera l and high diagonal associa t ion.
Resu lts for other d ata sets indicate tha t Japa n conforms
t
the low ge neral-hig h
diag ona l pa t te rn G oo dm an and H ou t 1993).
Ov er t ime, genera l associat ion has chang ed m ore than d iago nal associa
t ion in the Un i ted S ta tes Ho ut 1988) an d the N eth er lan ds Ga nzebo om an d
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 319/375
302
de Graaf 1984). In both countries the general association has declined by
more than 30 percent since the mid-1960s while diagonal association has
changed little if any.
Explanations for Cross National and Temporal
ifferences
The leading alternative to the Lipset-Zetterberg hypothesis of no
cross-national differences in mobility was the thesis of industrialization
(Treiman 1970). The main proposition of the thesis of industrialization
concerns the rationalizing effects of bureaucratic management. As firms
com pete for talent in a legal environm ent defined by the righ ts of citizen
job-seekers, they will be forced to assess workers on observable criteria:
This does not mean that family background no longer influences careers.
What it does mean is that superior status cannot any more be directly
inherited but m ust be legitimized by actual achievements that are socially
acknow ledged (Blau and Duncan
1967,
p.
430).
Blau and Duncan call this
process expand ing universalism . The thesis of industrialism posits that
expanding universalism will immutably erode the class barriers repre
sented by the association between origins and destinations.
6
From this
po int of view , social progress can
be
m easu red in the rate of decrease in the
association be tween origins and destina tions (Kerr 1959).
This emphasis on law-like, exogenous change, has given way in the 1990s
to a focus on political determ inan ts of mobility in general Nee
1991,
for state
socialist societies; Esping-Anderson 1992, for market economies) and the
association between origins and destinations in particular (Grusky and
Hauser
1984;
Erikson and G oldthorpe 1987,1992a; Ganzeboom et al . 1989).
The diffuse pressures brought on by rationalizing forces were insufficient,
in and of themselves, to foster equality of opp ortun ity. On the o ther hand,
intercep ting market forces through the pow er of the welfare state or state
socialism can be very effective in leveling social distinctions, including
distinctions based on origins. The politics of redistribution in Sweden and
Hungary have made very clear how the state has the potential to affect
equality of oppo rtun ity by affecting social mobility. The differences am ong
Hung ary, P oland, and Czechoslovakia caution against making too much of
a reified state socialist regimen (Erikson and Goldtho rpe 1992a, p . 373).
Anchoring the other end of the political spectrum , Ireland and France (and,
in its way, Japan) defend small, family-based enterprise in ways that
reinforce diagonal association. The German emphasis on industrial em
ployment and a stratification by skill in the w orking class reinforce the level
of association there (Erikson and Go ldthorpe 1987; Esp ing-Anderson 1992).
The American combination of rapidly changing op ppo rtunity in a con
text of little or no s tate activity is difficult to reconcile with the em erging
political theory. The federal state does virtually no redistribution from
class-to-class, yet the general effects decline (Hout 1988). The federal
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 320/375
Q
initiative
to
advance the
civil
rights of African A mericans actually increased
class differentiation within
the
black population (Wilson
1978;
Hout 1984b;
1986). Of course the federal government is not the only source of public
services, and, as I argue in the next section, the educational policies of
individual states, especially their support for their state universities, fos
tered mobility.
Public Higher ducation
Public policy can affect the association between origins and destina tions
either directly or indirectly. The most pervasive reform of the twentieth
century has been the worldw ide expansion of secondary and higher educa
tion (Me yeretal.
1982).
A mong its many consequences
is
the effect it has had
on mobility and life chances. s governm ent action, it is the exemplar of the
indirect policy instrument.
Everywhere wom en have been educated in large num bers, closing the
gap between the amount of education sons and daughters receive at the
very time when sons education was rising rapidly (Blossfeld and Shavit
1993).
Even when the reforms have little or no effect on the specific
transitions from school-to-school that make up the educational system , the
association tends to decline because origins are more important for
the
early
transitions than for the later ones (Mare 1980; Smith and Cheun g 1986;
Raftery and Hout
1985;
Shavit and Blossfeld 1993). In many cou ntries the
improved access to education leads to improved em ploym ent prospects as
well, reducing the association between socioeconomic origins and destina
tions in many countries (Ganzeboom et al. 1989).
The United States shows the potential for a successful approach that
operates indirectly through education (and without much coordination
am ong d ispara te policy-making jurisdictions). In part the Am erican case is
instructive precisely because educational expansion is the only attempt to
use pub lic action as a means of fostering equality of oppo rtunity . If several
policies were in place, it would not be so clear that educational expansion
was the effective policy. It is also clear that expansion was the operative
factor; little effort was expended to affect the other parameters of educa
tional stratification. C ontem porary deba tes stress cost as a factor because
the costs of higher education are rising so rapidly and so visibly. They m iss
the importance of access. When the cost of staying in school is the main
impediment to educational advancement, reducing costs through public
subsidy is a highly effective policy device for increasing equality of edu ca
tional opportunity and equality of educational outcome. Ho wev er, experi
ence has show n that cost
is
not always
the
main imped iment
to
educa tional
advancement.
Britain s famous reform of 1944 greatly reduced the cost of education to
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 321/375
3 4
stu de nts an d their families w itho ut redu cing the differences am on g classes
in their rates of education al adv anc em ent. In po st-w ar B ritain, the pre-ref o rm
system of pr ivate academic education could not meet the needs of the
growing middle c lass who wanted academic educat ions for the i r chi ldren
but could not afford them. Once the cost was reduced, the middle classes
imm edia te ly took adv anta ge them. The working c lasses did not respo nd as
quickly. The net result was that each class 's par t icipation in academic
education af ter 1944 was in proportion to i ts par t icipation pr ior to the
reforms (H alsey, He ath , and Ridg e 1980; Jonsson a nd Mil ls
1993;
Kerckhoff
an d Trott
1993).
Am on g the fac tors blunt in g w orkin g c lass par t ic ipa t ion are
the cul tu ra l bases of se lec tion em be dd ed in the tes ting ins t ru m ents
(Bo urd ieu 1977; O gb u 1987; Cla ncy 1986; Ly nch 1990) an d the sign als
(subtle and otherwise) f rom teachers that poor and working class s tudents
are not fit for acad em ic wo rk (Willis
1979;
MacLeod 1987). Poten tial ly m ore
re levant than these indirec t fac tors for unders tanding why reducing cos t
ha d so little effect on class differentials (but unr es ea rch ed as far as w e kn ow )
is the perceiv ed ra te of return to acad em ic educ atio n (Becker 1975). Ch i ld ren
of the affluent w o rk er s of the late 1950s an d 1960s (G old tho rpe et a l. 1969)
may have inferred a low rate of return to academic education from their
fa thers ' comb inat ion of low educat io n an d com for table incom e - a percep
tion l ikely to lead to un de rin ve stm en t in acad em ic edu catio n (Becker
1975).
Co sts are mo re directly l inked to incom e pe r se than to social class .
7
The
work linking costs to class barriers implicitly assumes that social classes
def ine d by o ccu pation h av e different levels of incom e. If every class ha s the
same economic s tanding ( income and weal th) on average , then changing
the cost of education may affect some individuals at the low end of their
class 's dis tr ib utio n of econ om ic s tand ing , b ut w ill no tb e l ikely to affect class
differences in educa tion al p ar t icipa tion rates . Few societies are at r isk of
incom e inequal it ies d isappea r ing. How ever , th is factor comp lica tes analy
ses of t rends in educat ional oppor tuni ty in socie t ies where the income
dis t r ibut ion h as change d. Most notably , the decrease in income inequal i ty
in Sw ede n in the twe nt ie th century in t rod uced amb igui t ies in to the in ter
preta tion of the changes ob served by Jonsson (1993; also Jons son a nd Mills
(1993).
Educational policy makers control the s ize of the public educational
system as well as the cost of at tending. Within broad bounds they can
directly affect educational outcomes as effectively by manipulating the
nu m be r of s tud ents ad mit ted as by changing the tu it ion. The expan s ion of
publ ic ly funded academic educat ion, especia l ly secondary edu cat ion , con
tr ibu ted at least as m uc h as low er tuit ion to equ ali ty of edu catio nal op po r
tuni ty in the twe nt ie th cen tury in man y countr ies (M are 1980; Smith an d
C h e u n g 1986; H o u t 1989; Shav it an d Blossfeld
1993;
Raftery a nd H o u t 1993).
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 322/375
3 5
FIGURE 13.6 Enrollments in Higher Education as a Percentage
of 18-22 Year Olds by T ype of Institution, United States, 1948-1992
80 T
70
60
-I
50
40 -
3 0 -
20
10 4
o
4-yr.
Public
Private
1940 1950 1960
1970 1980 1990 2000
Expansion creates opportunity by eliminating most forms of selection,
including class-based selection, at the lower levels of the educational
system. Equality of educational
outcomes
does no t necessarily follow, as the
expansion spreads not only through secondary schools but also through
universities and g radua te schools. Expansion works to reduce the associa
tion between
class
and educa tional outcom es because, in most societies, the
degree of class-based selection is greater at the lower than at the higher
rungs of the educational ladder (see Shavit and Blossfeld 1993 for a review
of the
evidence).
With the rungs that figure m ost in class-based selection no
longer a point of dep artur e from the system,
the
overall association between
class and outcome is typically reduced .
8
The sheer size of the expansion of American public higher education is
important to bear in mind first because greater size means higher cost and
second because the expansion was undertaken under extremely high
demographic pressure. The federal government provides funds to US
schools, but it administers only the military academies. State and local
governments administer the public schools; there are also many private
schools and universities (Coleman et al. 1982; Hout et al. 1993). Between
1945 and 1992, post-secondary enrollments as a proportion of American
youth 18-22 years old rose from
2
percent to
percent even as the size of
cohorts increased (see Figure 13.6).
9
Nearly all of that growth was in the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 323/375
3 6
FIGURE 13.7 Three Measures of the Share of Public Higher Education as a
Portion of all Higher Education - the ratio of enrollments in 4-year public
colleges and universities to total enrollments, the ratio of enrollm ents in
4-year public colleges and universities to enrollments in all public institutions,
and enrollments in private colleges and universities to enrollments in
4-year public colleges and universities, United States, 1963-1992
3 0
-I
1 1 1 1 1 1 1
196 1965 197 1975 198 1985 199 1995
public sector ; pr ivate college and universi ty enrollments have grown only
slightly faster tha n the eligible po pu la tio n (from 12 pe rce nt in 1965 of the
coh ort to 15 perce nt in 1992). M ean w hile en rollm ents in public four-year
colleges an d u niversi t ies as f raction of the college-age po pu lati on do ub led
(from 17 percent to 34 percent) between 1962 and 1992. Enrollment in
co m m un ity colleges and o ther two-year prog ram s accounts for a hu ge sha re
of the overall growth as total public enrollment raced ahead of enrollment
in pub lic colleges an d un iversi t ies with four-year d egre e pro gra m s. In 1963,
three-fourths of s tudents in public post-secondary insti tutions were in
four-year colleges an d univ ersi t ies; by 1975, four-year in sti tuti on s sha re
h a d fallen to jus t ove r half of the pu bli c enro llm en t (Figure 13.7).
The pe r iod of exp ans ion for public colleges and univ ersi t ies wa s 1952 to
1969.
In add it ion to acc om m od ating a r is ing fraction of each coho rt , pub lic
colleges and universi t ies share of total enrollment grew unti l the 1960s
when half of al l s tudents in higher education attended public colleges and
un ivers ities. D ur in g the first half of the 1970s, pu blic colleges an d un ive r
s i ties cont in ued to grow , bu t they could bare ly kee p up w ith cohor t gro w th
as ba by -bo om ers p ut pr essu re on their facili ties . Th en in the mid-1970s they
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 324/375
3 7
began to lag behind both community col leges and pr iva te univers i t ies .
Enrol lm ent ra t ios ha ve g one u p ag ain s ince 1984 thro ug h a comb inat ion of
sl ightly higher admissions, more f if th-year seniors , and smaller cohorts .
Pr iva te colleges and u nivers i t ies ha ve be gu n to ca tch u p w ith the publ ic
ones .
Their enro llm en t rat ios ha ve gro w n faster s ince the ear ly 1980s than
at any poin t betw ee n 1960 and 1980. Pr ior to 1955, pr iv ate colleges a nd
universi t ies enrolled more s tudents than the public sector did. By 1975, the
ratio of pr ivate- to-public ( four-year) enrollments had fallen below 50 per
cent (d ash ed l ine in Figure 13.7); i t reb ou nd ed to 55 perc ent by th e ear ly
1980s an d rem ains the re . Al l of which m ean s tha t , desp i te gro wth , pu bl ic
colleges an d un iversi t ies pla y a smaller role in US hig he r educ atio n in the
1990s th an they d id in the 1960s an d 1970s. Th ey ma y also be socially
sm aller in the sens e tha t rap idly r is ing in-state tuit ion for s tate un iversi t ies
in N ew York, M assach uset ts , M ichigan, and Cal ifornia rem ov e som e of the
dis t inc t ion between publ ic and pr iva te univers i t ies .
Impending demographic changes wi l l absorb any growth tha t may be
atte m pt ed in the nex t five yea rs an d for the f irst ten years of the nex t c en tury .
Th e nu m b er of 18 to 23 yea r olds will increase from 15 m illion in 1991 to 21
m illion in 2001 - a 40 percent increase in a decad e
(U.S.
Bureau of the C en su s
1994). Th at kind of gro w th led to a f lowering of the pub lic colleges an d
unive r s i ti e s be twee n
1966
an d
1975.
But in mo st
sta tes,
espec ially the larg est
ones ,
there is ne ither th e f inancial no r polit ical su pp or t for a secon d w av e
of growth. For example, cuts in s tate support have led California to cut
faculty instead of increas ing it wh ile po stp on ing p lans for a ne w U C c am pu s
that wa s s la ted to absorb the increase . Edu cat ional ne ed s wi l l hav e to
compete wi th manda ted pr i son expendi tu res in the nex t decade - com
p ou nd in g the pre ssu res on colleges an d univ ersi t ies in California. If the
pub lic insti tution s cann ot gro w , then the enrollm ent rates will ha ve to fal l.
H ow Am er ican highe r edu cat ion resp ond s wi ll a ffect the future of op po r
tun ity just as i ts exp ans ion pr io r to 1975 did.
The consequences of not meeting the needs of bigger cohorts will reach
bey on d the publ ic colleges and univers i t ies themselves . O pe nn ess in edu
cation is im po rtan t for the polit ics of mo bili ty because h igher edu cation —
and by inference
public
h igher educa t ion—promoted equa l i ty o f occupa
t ional opp or tu ni ty in the Un i ted S ta tes be tw een 1962 an d 1985. Contrac t ion
ma y po r tend the en d of tha t e ra . In doc um ent in g the
1962 85
im pr ov e m e n t s ,
I show ed tha t exp and ing h igher educa t ion—not exp and ing un ive r sa l i sm
— w as cr it ical (Ho ut 1988). I spec ulated that the gro w th of pub lic colleges
and univers i t ies was an especia l ly impor tant component of educat ional
exp ans ion , at least as i t related to the association be tw ee n social or igins an d
des tina tion s. If I w as r ight, then inequa li ty of occ upa tional op po rtu nit y
ought to remain at the level i t reached by the mid-1980s and may even
increase in the next decade.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 325/375
3 8
Severa l European countr ies have exper ienced enrol lment bot t lenecks
similar to the one about to jam American universi t ies . I reland, I taly, and
Russia al l experienced very rapid growth in their populations el igible to
enrol l in publ ic univers i t ies du e to a com binat ion of po pula t ion g row th and
an expansion of academic secondary education. In I reland the costs of
high er educ at ion a re born e by the s ta te . W ith no tui t ion m one y, e xpans ion
must be met out of general revenues (Clancy 1982) . The Ir ish Ministry of
Ed uca tion cou ld no t bu ild ne w facil it ies or hire ne w faculty fast eno ug h to
acc om m od ate a r is ing pool of el igibles that cam e of age be tw een 1970 and
1984. I r ish pol icy makers responded by a l lowing enrol lments to increase
only v ery s lowly— signif icantly s low er than the s ize of the el igible po ol w as
growing. Consequent ly , enrol lment ra tes as a propor t ion of academic
sec on da ry scho ol leavers fell from 1970 to 1984 (Clancy 1986; Raftery an d
H o ut 1985,1993). Th e available posit io ns w ere al located o n the basis of test
sco res , an d the his tor ical patte rn of no association betw een class or igins a nd
en try to univ ersi ty w as main tain ed (Raftery and H ou t 1993; Breen an d
Whelan 1992).
In Italy the enrollm ent cr isis w as me t by a laissez-faire a pp roa ch . A dm is
sion to univ ersi ty in I taly is by formula. Th e formula w as no t chan ged as the
po ol of el igibles grew , so a con stant f raction of that r is ing po ol w as ad m itte d
to un ivers i ty (Barbagli
1982).
Th e m on ey to con struc t ne w facilities or to hire
suff icient faculty was not al located. Consequently, s tudents admitted to
univ ers i ty found tha t they could not be adm it ted to courses they n eed ed.
D rop ou t ra tes—cal led ab an do nm en t in I taly—increased sharply- The
t ime to deg ree increased am on g those w ho d id not dro p out (Schizzerotto
1988),
and an association between class or igins and the probabil i ty of
gra du at in g f rom univ ers i ty em erged (Cobal ti and Schizzerot to 1993; Shav it
and Blossfeld 1993).
In Russia, too, a combination of larger cohorts and a higher rate of
qualifying for universi ty admission by graduating from an academic sec
on da ry school pu t pressure on the univers i t ies to adm it more s tu de nts . The
stu de nt fees w ere trivial , so the universi t ies had to w in app ro pri at i on s for
the ir gro w th from the centra l gov ernm ent . How ever , the g ov ern m en t ' s
pr io r i ty wa s univ ersal academic secon dary educ ation (af ter
1966;
see Co nno r
1991; Ge rber and H ou t 1995), leaving l i tt le m on ey for n ew h igh er ed uc a
tional facilities. Unlike Ireland, Russia had a history of sharp class differ
ences in univers i ty a t tendanc e . W hen the bot tleneck develo ped a t the po int
of entry to university, enrollment rates fell for all classes, but faster for
classes that tradit ionally had low enrollment rates than for the sons and
da ug hte rs of the in te llegencia (Gerber and H ou t
1995).
The net resul t w as
a significant increase in the class differentials in access to the universities
among cohorts born af ter 1960.
T h u s ,
we have three countr ies responding dif ferent ly to enrol lment
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 326/375
3 9
bottlenecks similar to the one American un iversities (especially public ones)
are about to face. They each responded differently. Ireland restricted
admissions but kept class differences from appearing. Italy deflected the
crisis from the university to the individual lecture and kept cl ss differences
from appearing at the point of entry while imposing stratification on
gradu ation p rospects. Russia rationed adm issions but because there w ere
significant cl ss differences to begin w ith, rationing increased ineq uality of
opportunity. Because tuitions and other costs are rising at public and
private universities in the United States, it seems unlikely that the Irish
mode l will em erge in the USA. The choice between the Italian and Russian
mode ls is neither pleasan t nor unfamiliar. A t many of the campuses of the
California State University system, frustration due to overcrowding and
increased time from matriculation to degree has been building. Rationing
adm ission and financing grow th by raising tuition has led to sha rp increases
in the family inco mes of freshmen at UC-Berkeley, UCLA, and the U niver
sity of Michigan (Fishlow 1993).
Direct Policy Approaches
Sweden has pursued the direct approach to fostering equality (Erikson
1983). The Sw edish p rog ram of full em ploym ent (much of it in the pu blic
sector),
high quality jobs, and progressive income redistribution has fos
tered equality of both educational and occupational opp ortun ity (Erikson
and Goldthorpe
1992;
Jonsson and Mills
1993;
Jonsson 1993). Full employ
m ent in the Swedish
c se
means not only low unem ploymen t but
lso
high
overall employm ent—alm ost 80 percent of Swedish a dults participate in
the labor force (Esp ing-Anderson 1992). The state pro vide s some pub lic
w ork s em ploym ent, bu t most state jobs are in health, education, and
welfare services. As such, they offer high w ages, steady em ploym ent, a nd
career opp ortu nities to state employees (Esping-Anderson 1992). Because
the state recruits persons from disadvantaged backgrounds to fill these
positions,
the
net resu lt is a lower association betw een origins and destina
tions than might otherwise be the case, and certainly lower than other
coun tries hav e achieved (Wong 1993).
The Netherlands undertook massive educational reform in the 1960s. It
redu ced the association between social origins and educational a ttainm ent
both by increasing
the ccess to
academic education for
ll
Dutch youth and
by low ering the effect of soci l class on the probability of enter ing academ ic
secondary school (DeGraaf and Ganzeboom 1993). In the process, it may
well have equalized the pattern of occupational mobility as educational
expansion did in the United States. However, the key change in the
Netherlands was the establishment of welfare state that rivals the Swedish
one in comprehensiveness. The combination dramatically reduced the
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 327/375
32
association between occupational origins and destinations in the Nether
lands between the mid-1960s and the mid-1980s.
State socialism w as expected to foster the same kinds of equalization of
opportunity as the Swedes and Dutch have achieved. European state
socialist states have not done so. Quotas for secondary and university
adm issions in Hungary increased work ing class enrollments, bu t did little
to redress the differences in chances between privileged and underprivi
leged s tudents (Szelenyi and A schaffenberg 1993). The evidence even
suggests an u ptu rn in inequality in the most recent
cohorts.
In Poland, the
places opened by educational expansion were filled by you ng w om en, not
the sons (or daughters) of the working class (Heyns and Bialecki 1993). In
Czechoslovakia, class was a weaker factor in school progress than else
where (except the Netherlands and Sweden) both before and during the
state socialist era. As we have seen, the enrollment
crisis in
Russia— brought
on by egalitarian secondary school expansion—led to
risin cl ss
barriers to
university enrollment (Gerber and Hout 1994).
Conclusion
The main point of this review has been to show the accumulation of
knowledge about social stratification that has accrued from mobility re
search since the early 1970s. Very little of that research has had a policy
focus,
yet the implications of our work for social policy are very importan t.
The clearest point is the last: sup porting higher education prom otes open
ness in society. The indirect effect of expanding public support for higher
education on equality of oppo rtunity in the United States has been a great
success story. The effect of class origins on class destinations has fallen by
between 30 and 50 percent since the early 1960s. Much of that change is
attributable to the expansion of public support for higher education. That
support has not continued to grow. In important states like New York,
Michigan, Texas, and , most recently, California, it has begun to shrink in
real terms. Building is deferred. Fees are hiked far faster than the m odest
rate of inflation.
The jokes during the 1992
U.S.
Presidential campaign about the choice
between m aking po tato chips and com puter chips is a real one. To say that
societies hoose the distribution of jobs overstates Esp ing-Anderson 's (1992)
argum ent somew hat, but the public choice of educational, emp loyment,
and welfare policies directly affects
the mix
of goo d and bad jobs in the
economy. Investment in human capital through public investment in
secondary and higher education is paramount.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 328/375
311
Notes
The UC-Berkeley Survey Research Center and Committeee on Research pro
vided funding. Thanks to Richard Arum and Gustavo Resendiz for research
assistance and to the conference participan ts for helpful com ments.
1.
T he one con troversial criterion tha t I will impose on materials I review here is
that they contain sufficient evidence of hierarchy in the data. By this I mean that to
be considered by m e,
a
class scheme must a dm it
to
substan tial variation in income,
prestige, and credentials (see Hout and Hauser 1992 for more discussion).
2. O rigin need only be a class that is occupied at some time in the
past;
destination
refers to current class. In this paper, I restrict my attention to intergenerational
mobility so origins refer to the class in which a person grew up . Significant research
on intragenerational mobility will feature in other presentations.
3.
The product of structural mobility and exchange mobility (along w ith p aram
eters that set the size of the samp le and the relative sizes of the origin classes) imply
the mo del of quasi-symm etry according to the formula:
F..-=a fl-8..
where F. is the
num ber of persons expected in cell (i, / ), a is the structural mobility multiplier for
class j,flfi. sets the sam ple size and the relative sizes of the origin classes, <5. =
8
are
the exchange m obility param eters for combinations
(i ,
j). See Sobel et
al.
(i98a) for
statistical details such as identification restrictions and estimation techniques. This
param eterization is referred to as the SHD formulation or the SHD m odel.
4. Since the Polish and Hungarian studies did not inquire whether self-employed
persons had employees the values for the two employer categories are missing not
zero in these two nations.
5. The US level reduc ed in the late 1970s and early 1980s (Hout
1988),
so using data
from 1982-85 wo uld conform to the CASMIN pa ttern quite closely.
6 The original Blau-Duncan formulation simply posited that more of the associa
tion wou ld become indirect via education (Blau and Duncan 1967, p. 430).
7.
An ind ivid ua l s or family s social class derives from th e source(s) of h is / h e r/
its incom e not from the amount of income itself Jencks
1991;
Erikson and Goldtho rpe
1992; Hout, Brooks, and Manza 1993).
8. Explanations for why class affects early selection more than later selection
differ. Mare (1993) review s the literature on this subject and presen ts a model for
testing some of the com peting theories using data on siblings (assuming that siblings
share the same class back ground). His results indicate that most if not all of the
familiar pattern of class effects being stronger at lower levels of education tha n at
higher levels is du e to the noncausal correlation betw een class and factors actually
being selected and not to a causal connection between class and survival.
9. These figures are not strict cohort measures because the requ isite data a re not
recorded. We have simply divided enrollments by the size of the population in an
app ropriate age g roup. The num erator includes all students, regardless of age. The
denom inator is the typical age range of post-secondary enrollmen t. Even this ratio
could no t be calculated w ithout som e manipu lation of the data at hand. Counts of
persons by single yea r of age are not available for most
years,
so we took the num ber
of 20-24 year olds tw o years later as our estimate of the num ber of 18-22 year o lds
in a given year Because the num erator covers just one year while the denom inator
covers five, the percents in figure 13.6 are 500 times the ratio of enrollments to
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 329/375
3 2
pe rso ns 18-22 years old.
10.
Ch ang es in the associat ion betw een o rigins and dest ina t ions w il l lag 10 to 15
years behind changes in higher educat ional pol icies because (a) the students have
to finish their degrees and establish themselves in careers while (b) the older, less
edu cate d cohor ts pas s out of the labor force.
11.
In California fees at the UC and CSU cam pus es ha ve gon e up 144 perc ent in
three years while the consumer price index has gone up 6.5 percent .
e f e r enc e s
Barbagli, Marzio. 1982. Educating for Unem ployment. N ew York: Colum bia U niver
si ty Press.
Becker, Gary. 1975.
Human Capital.
Chicago: Univers i ty of Chic ago Press.
Blau, Peter M. and O tis D udle y D unca n. 1967. The
Am erican Occupational Structure..
New York: Wiley.
Bo urdieu , Pierre . 1977. Reproduction in Education Society and Culture. Beverly Hills:
Sage.
Breen, Richard and C hristo phe r T. W helan. 1992. Explaining the Irish Pat tern of
Social Flu idity : Th e Role of the Po litical, P p. 129-151 in The Development of
Industrial Society in Ireland edi ted by J .H.Goldthorpe and C.T. Whelan. Oxford:
Oxford Universi ty Press.
. 1993. Fro m Ascript ion to Ac hievem ent? Or igins, Edu cat ion and Entry to
the Labour Force in the Republic of Ireland During the Twentieth Century.
Acta
Sociologica 36:3-17.
Breiger, Ronald, (forthcom ing). Socioeconomic Ac hievem ent and Social Struc
t u r e . Annual Review of Sociology
21
Clancy, Patrick. 1982. Participation in Higher Education. Dubl in : High er Educat ion
Authori ty .
Clancy, Patrick.
1986.
Social isat ion, Selection, an d Re prod uct ion in Ed uca t ion. Pp .
116-136 in
Ireland: A Sociological Profile
edi ted by Pat r ick Clancy , Shee lagh D rud y ,
Kathleen Lynch, and Liam O'Dowd. Dublin: Inst i tute of Public Administrat ion.
Clo gg, Clifford C 1982. Th e An alysis of A ssociation: M ode ls for Social D ata .
Journal of
the
American Statistical Association
71:803-815.
Cob al t i , A nton io an d Anto nio Schizzerot to . 1993. Inequ al i ty of Ed ucat ion al O p
po rtu ni ty in Italy. Pp . 155-176 in
Persistent Inequality: Changes in Educational
Opportunities in Thirteen Countries edi ted by Yossi Shavit and Hans-Peter
Blossfeld. Boulder, Co.: W estvie w P ress.
Co lem an, James S., Tho ma s Hoffer and Sal ly Kilgore.
1982. High School Achievement:
Public Catholic and Private Schools
Compared.
New York: Basic Books.
Connor, Walter D. 1991. The Accidental Proletariat: Workers Politics and C risis in
Gorbachev s
Russia.
Princeton: Princeton Universi ty Press.
De Graaf Paul M., and H arry B. G. G anze boo m . 1993. Fam ily Ba ckgr ound and
Edu cat ional Atta inm ent in the N ethe rlan ds for the 1891-1960 Birth Co ho rts . Pp .
76-100 in
Persistent Inequality: Changing Educational Attainment in Thirteen Coun-
tries edi ted by Yossi Shavit and Hans-Peter Blossfeld. Boulder, CO: Westview
Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 330/375
313
Erikson, Robert , John H. Goldthorpe, and Lucianne Portocarero. 1979.
Intergenerational Class Mobility in Three Western European Societies: En
gland, France, and Sweden.
British Journal
o f
Sociology
30:415-430.
Erikson, Robert.
1983.
Changes in Social Mobility in Industrial Nations: The Case
of Sw eden.
Research
in
Social Stratification and Mobility
2:165-195.
Erikson, Robert, and John H. Goldtho rpe. 1985. Are Am erican R ates of Social
Mobility Exceptionally High?
European Sociological Review 1:1 22.
. 1987. Com monality and Variation in Social Fluidity in Indu strial Nations
I.
European Sociological Review
3:54-77.
. 1992a.
The Constant
Flux:
Comparative Analysis
of
Social Mobility
in
Industrial
Nations. Oxford: Oxford University Press.
. 1992b. The CASMIN Project and the Am erican D ream.
European Sociologi-
cal Review
8:283-306.1992.
Esp ing-A nde rson, Gosta. 1992. Ch angin g Classes: Mobility Regimes in
Post-Industrial Econom ies. European University Institute. Unpublished m anu
script.
Featherman, David, F. Lancaster
Jones,
and Robert M. H auser. 1975. Assumptions
of Mobility Research in the United States: The Case of Occupational Status.
Social Science Research
4:329-360.
Featherman, David and R obert M. Hauser. 1978.
Opportunity and Change.
New York:
Academic Press.
Fishlow, Harriet. 1993.
Statistical Profile
of
Berkeley
Freshmen. Oakland, CA: Office of
the UC President.
Ganzeboom, H arry
B.
G., and P.M. De Graaf. 1984. Intergenerational Mobility in
the Nethe rland s in 1954 and 1977, in
Social Stratification
and Mo bility in the
Netherlands.
edited by B. Bakker,
J.
Dronkers and H.B.G. Ganzeboom. Amsterdam :
SISWO.
G anz ebo om , Ha rry B. G., Do nald J. Treim an, and Ru ud Luijkx. 1989.
Intergenerational Class Mobility in Comp arative Perspective .
Research in Social
Stratification and Mobility 8:
3-84.
Gerber, Theodore P., and M ichael Hout. 1995. Educational Stratification in Russia
During the Soviet Period.
American Journal o f Sociology
101:
forthcoming.
Go ldtho rpe, John H., David Lockwood, Frank B echhofer, and Jennifer Piatt. 1969.
The Affluent Worker in the Class Structure.
Cambridge: Cambridge University
Press.
Goodm an, Leo A. 1965. On the Multivariate Analysis of Three Dichotomous
Variables.
American Journal
o f
Sociology
71:290-301.
. 1970. The Multivariate Analysis of Qualitative Data: Interactions Among
M ultiple Classifications.
Journal of the American Statistical Association
65:226-256.
______ 1972. Some Multiplicative Models for the Analysis of Cross-Classified
Data. Pp . 649-696 in
Sixth Berkeley Symposium on Mathematical Statistics
edited
by Luciene Le Cam. Berkeley: University of California Press.
,
1979. Simple Models for
the
Analysis of Association in Cross-Classifications
Having Ordered Categories.
Journal
of the
American
Statistical
Association
74:537-552.
. 1984. The Analysis of Cross Classified Data Having Ordered Categories. Cam
bridge, M A: Ha rvard University Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 331/375
314
. 1991. M easur es, M odels, and Graph ical Displays in the Ana lysis of
Cross-Classified Data. Journal ofthe A merican Statistical Association 86:1085-1138.
G oo dm an , Leo A., and M ichael Ho ut . 1995. Stat ist ical M etho ds and G raph ical
Disp lays for the Analysis of Associat ion as it Varies Between Cou ntrie s, A m on g
Sub grou ps , and Over Tim e. Paper p resen ted a t the annu al me et ing of the
Am erican Socio log ica l Associa t ion , W ashington , D C , 18 A ugu st 1995 .
Gr usky , David, and Robert M. Ha user. 1984. Co m par at ive Social Mo bil i ty Revis
i ted. American Sociological Review 49:19-38.
Halsey, A.H., A. Heath, and J .M. Ridge. 1980. Origins and Destinations. Oxford:
Clarendon Press .
H aus er, Robert M., Peter J Dick inson, Ha rr y P. Tra vis, an d Jam es M. Koffel. 1975a.
Tem pora l Ch anges in Occupat ional Mobil ity am ong M en in the U.S .
American
Sociological
Review
40:279-297.
H aus er, Robert M., Peter J Dick inson, H ar ry P . Tra vis, and Jam es M. Koffel. 1975b.
St ruc tura l Chang es in Occupat ional M obi li ty am on g Men in the U .S . American
Sociological
Review
40:585-598.
Ha user , Rober t M.
1979.
Som e Explora tory M etho ds for A nalyz ing Mob il ity Tables
and Other Cross-Classified Data. Sociological
Methodology
11:141-158.
Ha user , Rober t M. , and D avid Grusky . 1988 . Occupa t ional Com posi t ion , Struc
tura l Mobil i ty , an d Cros s-Na tional V ariat ion in Mobil i ty . American
Sociological
Review 53:723-741.
H ey ns, Barbara, an d Ireneusz Bialecki. 1993. Edu cat ional Inequa l i t ies in Po stw ar
Po lan d. Pp . 303-336 in
Persistent Inequality: Changing Edu cational Attainment in
Thirteen Countries,
edi ted b y Yossi Shavit an d H ans-P eter Blossfeld. Bou lder,
CO :
Westv iew Press .
Hout , Michael . 1983. Mobility Tables. B everly Hills, CA : Sag e.
.1984. Sta tus , Au tonom y, and Tra in ing in Occupat ional Mobi l ity .
American
Journal of Sociology 89:1379-1409.
. 1984b. O ccu pa tion al Mobility of Black Men. American Sociological Review
49:308-322.
. 1986. O ppo r tun i tyan d the M inor i t yM idd le Cla ss : AC om par i so no f B lacks
in the United States with Cathol ics in Northern Ireland.
American
Sociological
Review
51:214-223.
. 1988. M ore U niversa l ism, Less St ruc tura l Mobi l ity : The Am erican O ccupa
tional Structure in the 1980s. Am erican Jou rnal of Sociology 93:1358-1400.
.
1989.
Following in
Father's Footsteps: Social Mo bility in
Ireland Cam br idge , MA :
Harvard Univers i ty Press .
H ou t , M ichael , an d Robert M. Ha user. 1992. Sy m m etry and Hie rarch y in Social
Mobil i ty .
European
Sociological
Review
8:239-265.
H ou t , Michael , A dria n E. Raftery, and E leanor O. Bell . 1993. M akin g the G rad e:
Ed ucation al Stratification in the United States, 1925-1989. P p. 25-50 in P ersistent
Inequality: Changing
ducational
Attainment in Thirteen Countries,
edi ted by Yossi
Shavit and Ha ns-P eter Blossfeld. Boulder, CO : W estview Press.
H ou t, Michael, Cle m B rooks, an d Jeff Ma nza . 1993. Th e Pers istence of Class es in
Post-Industrial Societ ies.
International Sociology
3:259-277.
Ishida, Hiroshi. 1993. Social
Mobility in Contemporary Japan.
Sta nfo rd, CA-.Stanford
Universi ty Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 332/375
315
Jencks,
Christopher. 1991.
Is
the Urban Underclass Increasing?" Chapter
2
in Urban
Underclass, edited by Christopher Jencks and Paul E. Peterson. Washington , D.C.:
Brookings Institution.
Jonsson, Jon 0 .1 9 9 3 . "Persisting Inequalities in Sweden." Pp. 101-132 in
Persistent
Inequality:
Changing Educational Attainment in Thirteen Coun tries,
edited by Yossi
Shavit and Hans-Peter Blossfeld. Boulder,
CO:
Westview Press.
Jonsson, J.O. and C. Mills. 1993. "Social Class and Educational Attainment in
Historical Perspective: A Swedish-English Comparison, Parts I and II."
British
Journal of Sociology
44:2-3.
Kerckhoff
Alan C , Richard T. Cam pbell, and Idee Winfield-Laird. 1985. "Social
Mobility in Great Britain and the United States." American Journal of Sociology
91:281-308.
Kerckhoff
Alan C , and Jerry M.
Trott.
1993.
"Educational Attainm ent
in a
C hanging
Educational System: The Case of England and W ales." P p. 133-154 in
Persistent
Inequality:
Changing Educational Attainment in Thirteen Countries,
edited by Yossi
Shavit and Hans-Peter Blossfeld. Boulder,
CO:
Westview Press.
Kerr, Clark, J.T. Dunlop, F.H. Harbison, and C.A. Myers. 1960.
Industrialism and
Industrial Man. C ambridge: Harvard University Press.
Lipset, Seym our M artin, and R einhard Bendix (eds). 1959.
Social Mobility in Indus-
trial
Society. Berkeley, CA: University of California Press.
Logan, John A. 1983. "A Multivariate Model for the Mobility Table."
American
Journal o f Sociology 89 (Septem ber): 324-349.
Lynch, Kathleen. 1990. "Reproduction: The Role of Cultural Factors and Educa
tional Mediators."
British Journal o f Sociology of Educat ion
11:3-20
MacLeod, Jay. 1987. Ain t No M akin It (First edition). Boulder: Westview Press.
Mare, Robert D. 1980. "Social Background and Educational Continuation Deci
sions." Journal of the American Statistical Association 75:295-305.
. 1993. "Educational Stratification on Observed and U nobserved Comp o
nents of Family Background." Pp. 351-376 in
Persistent
Inequality:
Changing
Educational
Attainment in Thirteen
Coun tries, edited by Yossi Shavit and Hans-Peter
Blossfeld. Boulder, CO: Westview Press.
Meyer, John
W.
and
W.
Richard
Scott.
1983. Organizational E nvironments: Ritual and
Rationality. Beverly Hills, CA: Sage.
Nee, Victor. 1991. "Between Redistribution and Markets in China." Am erican
Sociological Review
56:267-282.
Ogbu , John U. 1987. "Variability in Minority School Performance: A Problem in
Search of an Exp lanation."
Anthropology Education Quarterly
4:312-334.
Pontinen, Seppo. 1984. A
Comparison
o f
Mobility
in
Scandinavian
Countries. Helsinki
Sociological Institute.
Raftery, Adrian E., and Michael Hout. 1985. "Does Irish Education Approach the
Meritocratic Ideal?"
The Economic and Social Review
16 (January): 115-140.
. 1993. "Maximally Maintained Inequality: Expansion, Reform, and Oppor
tunity in Irish Education, 1921-1975." Sociology o f Education 66:41-62.
Roos,
Patricia. 1986.
Gender and
Work:
A
Comparative Analysis
of
Industrial Societies.
New York: State Un iversity of New York Press.
Schizzerotto, A. 1988. "II ruolo dell'istruzione nei process di mobilita." Polis
1:83-123.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 333/375
326
Shavit, Yossi, and Hans-Peter Blossfeld, (eds.) 1993.
Persistent Inequality: Changing
Educational Attainment in Thirteen Countries. Boulder, CO: Westview Press.
Sobel , M ichael E., Michael Hou t , and O tis D udley D unca n. 1985. Exc hang e,
St ruc ture , and S ym me try in Occupat ional M obi li ty .
Am erican journal ofSociology
91:359-372.
Sorensen, Jesper B. 1992. Locat ing Class Cleavage s in Inter-gen erat ional Mo bil i ty:
Cross-National Commonali t ies and Variat ions in Mobil i ty Pat terns.
European
Sociological
Review 8:267-279.
Sm ith, H erb ert L. an d P.P. L. C he un g. 1986. Tr en ds in the Effects of Fam ily
Background on Educat ional Attainment in the Phi l ippines. American Journal of
Sociology
91:1387-1408.
Sze lenyi, Szonja, an d K are n Asch affenberg. 1993. Ineq ualitie s in Edu cati ona l
O ppo rtun i ty in H ung ary . Pp . 273-302 in
Persistent Inequality; Changing Educa-
tional Attainment in Thirteen Countries,
edi ted by Yossi Shavit and Hans-Peter
Blossfeld. Boulder, CO: Westview Press.
Tre im an, Donald
J
1970. Ind ustr ializ ation and Social Stratification. Pp . 207-234 in
Social Stratification: R esearch and Theory for the 1970s,
ed i ted by Edw ard O.
Laumann. Indianapolis: Bobbs-Merri l l .
Willis, Paul. 1979.
Learning to Labour
London : Saxon H ouse .
W ilson, W il liam
Jul ius.
1978. The Declining Significance of Race. Chicago: U nivers i ty
of Ch icago P ress,
W ong, Ray m ond Sin-Kwok. 1990. U nde rs tand ing Cross-Nat ional Varia t ion in
Occ upat ional M obi l ity .
American Sociological Review
55:560-73.
. 1992 Vertical an d N onv ertica l Effects in Class M obility: C ros s-N atio nal
Varia t ions . American
Sociological
Review 57:396-410.
.
1993.
Po stw ar Mobil i ty Tre nds in A dva nce d Industrial Societ ies.
esearch
in Social Stratification and Mob ility
12:121-144.
W ong, Raym ond Sin-Kwok, and Robert M. Ha user . 1992. Trend s in O ccupat ional
Mobi l i ty in Hun gar y U nde r Socia l ism. Social Science
esearch
21:419-444.
Wright, Erik Olin. 1985.
Classes.
London: Verso.
. 1987. Reth inking, O nce Aga in, The Co ncept of Class Stru ctur e. Pp .
269-348 in
The Debateon Classes,
edi ted by Erik Olin Wrigh t , U w e Becker, Joha nna
Brenner , Michael Burawoy, Val Burr is , Gug l ie lmo Carchedi , Go rdon M arshal l ,
Peter F. M eiksins, Da vid R ose, A rth ur St inchcom be, an d P hil ippe V an Pari js .
Lond on: Verso .
Xie,
Yu. 1992. The Log-M ult iplicat ive Layer Effect Mod el For Co m pa rin g Mo bil i ty
Tables . American
Sociological
Review 57:380-395.
Ya m aguc hi , Ka zuo. 1983. The Struc ture of Interg enera t ional Oc cupa t ional Mobil
i ty: Ge neral i ty and Specifici ty in Resource s, Ch ann els, and Barriers . American
journal of Sociology
88:718-745.
. 1987. M odels for Mobil i ty Tables: To w ard P arsim ony and Sub stanc e.
American Sociological Review
52:482-494.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 334/375
4
Stratification and Attainment
in a Large Japanese Firm
Seymour Spilerman
olumbia University
Hiroshi Ishida
University
of okyo
Introduction
Japanese management pol icy has been a source of keen fasc inat ion to
social scientis ts in the United States . M an y issues relat ing to the org aniz a
t ion of eff icient, h igh qua li ty prod uc tion an d de livery of services ha ve b ee n
resolv ed in a m an ne r that is believed to be quite different f rom U .S.
formula t ions . Japan ese indu s try , for exam ple , has ins t i tu ted prac t ices su ch
as qu alit y con trol circles, just-in-tim e delive ry, affiliated satellite firms, a n d
contrac t ing out ar rangements—adminis t ra t ive s t ruc tures tha t have few
precedents in the United States .
The human resource sys tems of la rge companies in Japan are a lso
characterized by practices quite different from those in U.S. f irms, both in
the assumed employment re la t ion and in the approach taken to career
deve lop me nt . Not ions of l ife -t ime emp loym ent , h i r ing pe rm ane nt wo rk
ers directly f rom school, the rar i ty of lateral entry, the senior i ty reward
system (nenko) , and salary compression are some of the features that are
dis t inc t ive to perso nne l man age m en t in Japan (Cole
1979;
Ko ike 1988). T his
is no t to sugge st th at the se practices are entirely absen t f rom U.S. f irms, b ut
i t is rare for several to appear in the same company, though IBM in past
dec ade s may ha ve been an exception (Foy 1975) . In large Japa nese co m pa
nies
1
these features tend to be components of an integrated personnel
system, possibly reinforced by cultural norms that s tress social harmony,
pr imacy of corpora te goals over individual aspira t ions , and deference to
e lders (Dore 1973, pp . 51-52,297-298; Ko ike 1988, p p . 4-7; Ro lan d 1988, p p .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 335/375
3 8
72-75; b u t also see Lincoln a nd Kalleberg 1990).
W hen ex amining the hu m an resource sys tem of a Japanese f irm o ne can
read ily obse rve how so me of the com po ne nt features are knit tog ether and
w or k in tan de m , as well as identify s trains in an othe rwise co heren t set of
personnel practices . Contracting-out arrangements , for example, facil i tate
a policy of l ifet ime employment by permitt ing f luctuations in product
d em a n d to be pa ssed on to an affiliated satellite firm (Dore 1973, p . 39). A n
ear ly mandatory re t i rement age—55 to 60 is the common range—often
followed by rehir ing the sup eran nu ated wo rker (a t a lower wage) pro vid es
additional f lexibility to a firm in adjusting to the economic cycle, since the
rehired w orker is no longer protec ted by the l ife time com m itme nt .
A decis ion to offer l ifet ime employment carr ies several implications.
Firs t , personnel selection must be done with great care s ince a s trategy of
disp osin g of w ork ers wh o later reveal them selves to be po or perfo rme rs is
no t avai lable to m ana gem ent . Second, w he re l ifet ime emp loy m en t is
cou pled w ith a l inkage of salary to senior i ty, w ork er motiv ation a nd w ag e
cos t con ta inm ent can becom e prob lems s ince m an age m ent s denied the use
of som e po ten t beh avio ral reinforcers . Third , al th ou gh lateral ent ry is no t
formally foreclosed by a policy of l ifetime emp loy m en t, w he n m an y f irms
follow this practice few m id-lev el job cha ng ers are likely to be av ailab le for
recrui tm ent . These cons idera t ions heighten the imp or tance of t ra ining an d
socialization. The skil ls that will be req uired by a co m pan y ha ve to be
dev elop ed internally ; mo reover , fu ture corpora te leaders mu st be se lec ted
and groo m ed f rom the youn g recrui ts (Peck an d Tam ura
1976;
O E C D 1 9 7 3 ;
Cole 1979, pp . 40-42). This circum stance help s explain the lo ng-te rm
relationships with par t icular schools that are of ten sought by employers ,
s ince the educational inst i tutions can evaluate s tudents on the basis of
several years of observation (Rosenbaum and Kariya 1989) .
Tens ions ar ise because som e personnel prac tices do not mes h w el l w i th
oth ers . For exam ple, the socialization of m ana geria l recruits is inte nd ed to
instil l co m m itm en t to the f irm an d reinforce coho rt bo nd in g (Dore 1973, pp .
46-54).
Identification w ith the firm, rather than w ith a sub un it , is pro m ot ed
by a pol icy of rota t ing employ ees am on g occupat ional tasks and organiza
t ional uni ts ; in prac t ice , enco uraging a genera l is t or ienta t ion (H irono
1969, pp . 260-61). Ac cor din g to Ballon (1969a, p. 26), job rotation is essential
for adv ance m ent to senior m ana gem ent . Yet, la rge f irms mu st a lso d eve lop
specialis ts , and economic rat ionali ty suggests that such individuals be
en co ura ge d to w ork w ithin the do m ain of their expertise. W e kn ow lit tle
about how specialization is motivated or rewarded in a work context that
is largely geared to promoting a generalis t or ientation, though see Koike
(1991) and Pucik (1964a) for some insights into the career implications of
specialization.
2
Similar ly, the selection of em plo yee s for adv an cem en t to senior posit ion s
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 336/375
3 9
enta i ls a process of d i f ferentia tion am on g cohor t me mb ers— a problem at ic
activity in a contex t in wh ich cohesion is pr ize d. The re is som e c on sen sus
to the effect that these conflict ing objectives are managed by segregating
the m in t ime. Solidar i ty is em ph asiz ed d ur in g the f irst deca de or so of
employment , fac i l i ta ted by a pol icy of automat ic promotion through the
junior ranks. Only af terwards does differentiat ion begin in earnest , with
som e emplo yees earm arke d for h igh adm inis t ra t ive ran ks (Clark 1979, p .
112-119; Yoshino and Lifson 1986, pp. 146-147).
H ow this process unfolds is no t clear ly un dersto od . Fo llowing Ro senba um
(1984),
Pucik
(1985,
p.77) con ten ds that after the firs t deca de of em plo ym en t
a t ta inmen t can be desc r ibed by a tourna me nt mo de l , in wh ich employe es
who have advanced rapidly f rom one rank have sharply higher prospects
for swif t prom ot ion in the subs eq ue nt rank (see H an ad a 1987 for a s imilar
assessm ent) . In R ose nb au m 's formulation (1984, pp . 61-62), ear ly w inn ers
— me asured by r ap id p rom ot ion in the in it ia l r anks— have an opp or tun i ty
to co m pe te for high s tatus es, w hile losers can com pete o nly for lesser jobs.
A der ivative feature of the tournament model of careers is i ts weakness in
acco m mo dat in g er rors of exclus ion. I t lacks a me chan ism for br in gin g
early losers back into the com peti t ion for senior posit io ns (Ro senb aum 1984,
p .
288).
Ballon has suggested a very different decis ion calculus, one in which
product ivi ty in the ear ly ranks is not a dominat ing cons idera t ion in la ter
prom otio n decis ions . The ques t ion for Ballon is w ha t i s me ant by p rod uc
tivity? [ I]n Jap an econo mic perfo rma nce is no t so m uc h a m atte r of
ind ivi du al em plo ye es as [it is] of an entire organ izatio n (Ballon 1969a,
p.26).
Prom otion , in turn, is less t ied to the details of w ork p erform anc e tha n
to pro pe r organ iza t ional va lues . These are expected to m atu re wi th age
and tend to reveal themselves only in late career s tages. Assuming, then,
that rapid advancement in the ear ly ranks ref lects individual product ivi ty ,
one pred ictio n from Ballon's thesis is that s low init ial mobili ty m ay n ot b e
a handicap in la ter promotion decis ions .
3
Th ere is also the relate d issu e of the onsequen e of advancement a nd how
the effects of loss in the career com pet ition a re m an ag ed . This is a serio us
issue in a context of l ife time emp loy m en t because a wo rker wh o has been
pas sed over for pro m otio n m us t s ti ll be moti vate d to perform effectively in
lower ranked posit ions. The fact that the senior i ty component of salary is
often large in Japanese firms (Cole 1979,
p.41;
O ko chi, et. al. 1974, p p . 499-
500) operates to mitigate the problem, s ince this arrangement l imits the
m ater ial effects of loss in the prom ot ion co m pe tition . Cla rk (1979, p . 122)
also descr ibes a set of t i tles acco rded to em plo yee s on the basis of sen ior i ty
— honorable consola t ion pr izes —but which lack the author i ty tha t de
r ives f rom high rank in the pr incipal s tatus hierarchy of the f irm.
In th is pa pe r w e examine the re la ted them es of s ta tus adva nce m ent an d
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 337/375
32
com pens at ion level in a la rge Japanese com pan y in the mid-1990 's . W e
org aniz e the analysis aro un d the following specif ic issues: (a) W ha t are the
de term in an ts of pro m otio n and ho w do they vary by ran k in the f irm? (b)
What is the consequence of status attainment (relative to the effect of
senior i ty) for rem un erat ion level? (c) W hat kind of con ceptu al im age ry—
e.g., Ro senb aum 's tourna m ent mo del ; Bal lon 's formula t ion— best descr ibes
the way tha t oppor tuni ty and advancement are s t ruc tured for employees?
The information for this s tud y com es from the em plo yee d ata base of one
of the ten largest f inancial service co m pa nie s in Jap an . O u r da ta set is
unusual ly r ich in tha t we have complete work his tor ies of employees
cu rren t as of 1993, wh ich permits personn el i ssues to be addre ssed in som e
deta i l us ing the m eth od s of survival analys is . M oreover , whi le w e hav e
em ph as ize d career features that are fairly dis t inctive to Japa n, there h av e
been suggestions of change and a trend toward adoption of practices
com m on in wes te rn countr ie s—in bo th sala ry de te rmina t ion an d pro m o
tion policy (Marsh an d M an na ri 1976, p .
120;
Yo shino a nd Lifson 1986, p p .
152-155;
W olferen and M ur ph y 1994, p. IV-13). Because the f irm w e h av e
studied is considered by its officers to be fairly typical of large companies
( in the f inancial service sector) in its h u m an resourc e practices , by exa m in
ing the extent to which our f indings deviate f rom ear l ier accounts of
Japan ese wo rk sys tem s w e can assess the eviden ce for chan ge an d conv er
gence to wes tern ar rangements .
The S t r a t i f i ca t ion Sys tem for Manager ia l Employees
Before turnin g to the analys is of rank adva ncem ent an d sa lary de term i
na tio n, it is useful to ou tline the ma in features of the stratif ication sys tem of
the Japa nese f irm. Also, for a reference s tan da rd, we com pa re the Japa nese
co m pa ny w ith a large financial service organiz ation in the U.S. , w hic h h as
been stu die d in som e detail (Spilerman an d Lu nd e 1992; Peterse n an d
Sp ilerm an 1990).
An a logous to the U.S. f irm, the Japa nese co m pa ny is div ide d in to cler ical
an d m an ag eria l specialt ies . H ow ev er , (1) al l clerical w ork ers are w om en
and a lmost a l l manager ia l employees are men.
4
In the U.S. firm, by
comp ar i son , w om en cons t itu te 91% of cler ical w ork ers an d 46% of a dm in
is trat ive em plo yee s. (2) In the Japa nese com pa ny th ere s no mobi l i ty across
the c ler ica l /manager ia l d ivide; recrui ts to the manager ia l ranks come
direc tly from college. In the U.S. f irm, in con trast, there are ex tens ive
po stin g an d bi dd in g prov ision s to facili tate job transfers, an d these ha ve
resulted in approximately half of managerial entrants coming from the
cler ical ranks of the company.
Reg arding en trants f rom the external labor m arket in to the m anag er ia l
ranks ,
in the U.S . f irm so me 5 5% hav e had pr ior w ork exper ience ,
5
a
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 338/375
321
background that is rare in the Japanese com pany. In the U.S., new m anage
rial hires exhibit differences in level of educational attainm ent, w hich is not
th e
case
in Japan w here all entran ts hav e a college degree.
3)
In Japan, there
is a distinction between non-perm anent and perm anent wo rkers,
which is virtually coterminous with the clerical/managerial dichotomy.
Only permanent employees are promised lifetime employment.
To summ arize,
in
the Japanese firm m anagerial emp loyees are male; they
hav e been recruited directly from college; they com prise an entering cohort
that is homogeneous in age and in education; and most expect to spend their
full workin g lives within the com pany . Clerical em ployees, in comp arison,
are female, usually without
a
college degree. They
are
not recruited into the
managerial ranks and their tenure with the firm normally ends with
m arriage or childbirth. The significance of this distinction in the pe rson nel
categories is conveyed by the fact that the employee data base of the
Japanese com pany contains wo rk histories of only managerial em ployees.
Hence, our investigation is restricted to managerial employees and to men .
It shou ld
also
be emphasized that the sort of analysis a stud ent of careers
can do with the Japanese materials is less rich than what is commonly
un dertak en w ith U.S. data. In the study of the U.S. financial service
com pan y, for example, mu ch of the research effort w as devo ted to exam in
ing the effects on attainment of gend er, years of schooling, entry age, and
entry po rtal (prom otion from the clerical grades versu s entry by a new hire
directly into the man agerial ranks). In the Japanese company , how ever,
there is little variation on any of these variables. Indeed , this tendency to
hom ogeneity of an entry cohort— a frustration to the researcher—com bined
with the practice of encou raging a generalist orientation , facilitates the
comm on corporate policy of rew arding employees principally on the basis
of seniority. If there is little differentiation on
other
human
c pit l
variables the
significance of seniority as a determinant of productivity and performance is
heightened
The structure of rewards
In large work places in the
U.S.
the reward structure is commonly based
on the principles of modern compensation design (e.g., Wallace and Fay
1983;
Sibson 1981). The essential features of such hu m an resource systems
are a set of ranked salary grades (there are 20 in the U.S. company), the
slotting of job titles into the grade levels on the basis of a job evaluation
procedure, and a definition of promotion as an upward movement in the
salary grad e stru cture, rather than in terms of a chang e in job title. As
remarked elsewhere (Spilerman and Petersen 1994), the attractiveness of
this arrangem ent is that it frees m anag em ent from the tyranny of a techno
logically determ ined occupational distribution that constrains prom otion
allocations to the presence of vacancies. In the salary grade form ulation
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 339/375
322
management can more eas i ly award promotions on the bas is of mer i t .
Lacking a vacancy, a job title can either be reclassified into a higher grade
or the number of grades associated with the posit ion can be increased.
In com par ison, the reward s t ruc ture in the Japanese com pa ny h as three
dist inct d im ensio ns. First , there is a system of status ranks (often re ferred to
as s ta nd ard ran ks [e.g. , Clark 1979, p . 104]), a hiera rchy w hic h lacks a
cou nterp ar t in U.S. indu s try . The s tan dard ranks— ti tles such as depa r t
ment head, sec t ion head, sub-sect ion head—adhere to the individual and
are no t necessar i ly descr iptive of the job he doe s. The s tan da rd ran ks a re
fair ly un ive rsal in Jap an; as a result the t it les conv ey me an in g th ro ug ho ut
the society about an individual 's s tatus and provide a basis for social
co m pa rison s am on g emp loye es of different f irms. In the U.S. the closest
parallel is s tatus in the mili tary, in which rank conveys authority and
patterns deference relationships but is not indicative of an officer 's job
ass ignment .
The second dimension relates to functional respon sibility in the com
pa ny — on e 's ma nag er ia l author i ty . W hile there is a cor res pon den ce be
tween this dimension and standard rank, they are not formally identical .
Cla rk (1979, p p . 111-115), for exam ple , desc ribes an orga niz atio na l settin g
in which senior level supervisory posit ions are sometimes lef t unf i l led, in
order to permit capable junior employees to take on dut ies normal ly
associated with high s tandard rank, which they lack the tenure to acquire.
As Dore (1973, p . 68) no tes, [ t]he ad va nt ag e of this f lexible sys tem is tha t
i t a l low s faithful service by m en of me dioc re abil ity to be rew ar de d by an
increase in rank with ou t the disad van tage of dysfunct ional ly pro m otin g
them to posit ions of greater authority.
The thi rd dimens ion concerns
salary grade
The U.S. f inancial service
company uti l izes a set of 20 grades, essential ly a system of overlapping
salary range s . In the Japanese com pan y whic h we exam ined there are 34
gr ad es , each of w hic h specifies a ba se salary rate that is adju sted for
sen iority . M ore con seq ue ntia l is the different significance of the sala ry
gra de hierarc hy in the tw o cou ntr ies . As no ted ear l ier , in the U.S. c om pa ny
salary gra de cons t i tu tes a unified rew ard dimen sion. There are n o per
son al s tatu s rank s; also, the job t i tles—w hich convey functional respo nsi
bil i ties—are ma pp ed onto the salary gra de s. P rom otio n, in turn , is def ined
as m ov em en t in th is grad e hierarchy.
In contrast , in the tr ipar t i te divis ion of the Japa nese co m pa ny , s ta nd ard
rank c onsti tu tes the central factor in the allocation of s tatus , au tho rity, an d
career rewards (Rohlen 1974; p . 25). A lso, pr om ot io n is defined in term s of
m ov em en t w ithin this hierarchy. W hile salary gra de is correlated w ith
standard rank, the former is often tied to seniority, especially at the
be gin nin g of the career (Dore
1973,
p . 68; Ro hlen 1974, p.156). Ind ee d, th is
em ph asis on senior i ty in the sett ing of salary— rather th an job a ssignm ent—
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 340/375
323
is rational in an environment in which job rotation is encouraged and
em ploym ent is long-term. While a worker's wages might not reflect his
specific
job
duties at a given time point, the compensation model—as well
as notions of equity and em ployee expectations—is formulated in terms of
lifetime c reer
rewards not statically (Yoshino and Lifson
1986,
p . 152; Koike
1988, p . 134).
Since the system of stand ard ranks is the principal dim ension of stratifi
cation in the Japanese firm, w e have focused our investigation on mobility
within this status hierarchy . Because of the complex relation between
compensation and status rank—high remuneration is both a reward for
attainment and, possibly, a consolation prize as well for passed-over
employees—we also examine the consequence of rank attainment for
com pensation level.
The ttainment Process in the Firm
In Table
14 1
w e present cross-sectional information on the distribution
of personnel in the managem ent ranks in 1993 (column 1). From the
du ratio n figures (column 3) it is evident tha t a long period is spen t in the
non-management or trainee
6
status and that there is little variation in
dura tion in this rank (column
4,
row 1). This finding is consistent w ith the
reports of other investigators (e.g., Yoshino and Lifson 1986, 146; Pucik
1985,
p . 74), wh o hav e noted that the trainee period lasts from 10-15 years
and that prom otion is autom atic after a fixed interval in this status. (We
shall, how ever, have more to say about this assessment.) The figures in
colum n (5) provide rough evidence for the stability of the organization in
size and in status d istribution, indicating that turnove r in ranks
20 to 40
has
been fairly constant at about 110 pe rso ns /yea r. In ranks 20 and 30 this
turnover consists largely of promotions; however, many of the exits from
rank 40 are retirements and hence the decline in turnover in the highest
positions,
Which variables predict to prom otion? This issue is add ressed in Table
14 2
using Cox's proportiona l hazard m odel (Blossfeld, Hamerle and M ayer
1989,
chap. 3) with career history da ta from the 1961-82 en try
cohorts.
The
Cox model specifies that
h(tlX) = h
0
(t)exp(X'E) (1)
where >£' is a vector of covariates,
B_
is the vector of respective regression
coefficients, t is the w aiting time to prom otion, and h
0
(t) is an unspecified
base rate that is the sam e for all indiv idua ls. To perm it the possibility that
the determinants vary with level in the organization, we exam ined p rom o
tions from ranks 20, 30, and 40 separately. Thus, the regressions in each
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 341/375
TABLE 14.1 Distr ibutio n of Perso nnel by
Sta tus Rank in the
Japanese F inancia l Service Co mp any , 1993
(1)
Status
Rank
b
N a m e N
20 Tra in ee/N on-M anag em ent 1 ,106
30 Sub-Section Chief 542
40 Section Chief 921
50 Sub-Dept .Head 110
60 Dep art . Head 146
70 Director 5
N 2,830
(2)
Mea n Age of
Entry into
Rank (years)
23.05
34.04
38.89
47.71
49.61
—
(3)
Me a n
Dur a t i on
in Rank
c
(years)
11.03
4.65
8.46
2.8
1.75*
—
(4)
C.V.
d
of
Durat ion
.067
.260
.225
.421
r
—
(5)
N /
Me a n
Durat ion
100
117
109
39
—
Omit ted f rom the s tud y are female emp loyees , la tera l t ransfers f rom oth er compa nies , and emplo yees
wi thout a col lege educat ion . Exclus ions for these reasons am oun t to
ss than 5% of the firm's labo r force.
• Some min or rank s (in 20's) and rank s with sm all N 's (45, 80, 90) hav e been gro up ed
major rank.
c
Based on comple ted dura t ions .
Coeff ic ient of var ia t ion = s.d . of d ura t io n/m ean dura t ion .
Mean dura t ion in rank 60 is based o n current inc um bants , not comp le ted spel l s .
Va lue i s not comp arable to o thers in the colum n.
See note e.
w i t h a n a pp r op r i a t e
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 342/375
325
panel are based on spells in the particular rank. An employee can contribute
only one spell to a panel; however, if he has progressed through several
ranks he can contribute a spell to more than one panel. In each model, the
regresso rs predict to the rate of prom otion from the noted rank; incom plete
spells as of July 1993—the data collection date—are treated as censored
observations.
For reasons ou tlined earlier— the homogeneity of m anagerial recruits in
terms of educational attainment, gender, and (absence of) prior work
experience -
the
available regressors are few. Nonetheless, some distinctive
patterns emerge. With respect to advancem ent from the non-m anage
m ent status (Panel
A),
neither
age
at hire
on
which there
is a
range of some
3 years) no r college major predicts to prom otion . However, the models in
colum ns (3) and
4)
indicate
a
negative association between size of an en try
cohort and the promo tion rate, which would suggest a corporate policy of
insulating the higher ranks from annual variations in magnitude of the
intake. This finding w ould seem at variance with the contention tha t
prom otion from rank 20 is autom atic; however, we will shortly make
clear that even at this early career point the company has begun to make
distinctions among employees as well as to fine-tune the advancement
regime.
In the analysis of promotion from rank 30 (Panel B), we introduce a
regressor for dura tion in the prior rank. This term is intended to assess
whether individuals w ho have previously advanced rapidly are advan taged
with respect to current promotion prospects. The significant nega tive
coefficient that we find—long prior duration reduces the prom otion rate -
supports
this
possibility (column
1).
One explanation for the finding would
em phasize the sorting of workers on
the
basis of either tournam ent success
or
ability (unm easured
in
our data
set);
however, these initial results are also
consistent w ith a tracking or gate-keeping explanation.
Further insight into the mechanics of the attainment process can be
obtained from an analysis of promotion in rank
40
(Panel
C).
These results
provide support for a tracking type of explana tion. In particu lar, while
we continue to find a significant negative coefficient for time in rank 20,
there is no effect of du ration in rank
30,
the prior s tatus level (column I).
7
This result is not consistent with an ability sorting thesis or with a
tourna m ent m odel; both would suggest that recent job performance—
indexed here by duration in prior rank—should have greater impact on
prom otion chances than less proximate performance m easures. Instead,
our results suggest a process in which a critical evaluation is made early in
an employee's career and it is this decision, rather than later job perfor
mance, which determines the worker's subsequent promotion prospects.
Finally, we note from colum ns
3)
and (4) of the several panels that en try
cohort size has no impact on the prom otion rate after the trainee years (rank
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 343/375
3 6
TABLE 14.2 Pro m otion Regressions, by Status Rank in the Com pa ny
3
PANEL A. Promotion from Rank 20
Variable
(1)
2) 3)
e
4)
A ge
c
(years)
College Major:
d
B u s i n e s s / L a w
Science
Cohort Size
-2LL =
N =
Percent censored =
.0114
19296.60
1486
0.5%
.0020
-.0126
.1353
19278.37
1485
0.5%
.0726
-.0012*
19827.53
1486
0.5%
.0130
-.0061**
19287.47
1486
0.5%
Variable
b
PA NE L B. Pro m otion from Rank 30
(1) (2) (3)*
4)
A ge
c
(years)
Durat ion in Rank 20
College Major:
d
B u s i n e s s / L a w
Science
Cohort Size
-2LL =
N =
Percent censored =
(years)
.0013
-.1679*
12397.41
1119
11%
.0233
-.1925**
.3697*
.1405
12373.84
1118
11%
.0046
-.1667*
.0005
12415.88
1119
11%
.0001
-.1681*
.0010
12397.18
1119
11%
Variable
PAN EL C. Prom ot ion f rom Rank 40
(1) (2)
3)»
4)
A ge
c
(years)
Durat ion in Rank 20
Durat ion in Rank 30
College Major: '
B u s i n e s s / L a w
Cohort Size
-2LL =
N =
Percent censored =
(years)
(years)
.1605
-.4770**
-.0567
1830.74
300
37%
.1959
-.5114**
-.0552
.6714*
1812.10
299
37%
.1392
-.4981**
-.0344
-.0006
1834.75
300
37%
.1632
-.4890**
-.0610
-.0081
1830.23
300
37%
*p<.05, **p<.01
Regressions are Cox prop ort ion al haza rd m ode ls. Clock for the m ode ls is
du rat i on in rank. Data a re from 1961-82 entry c ohor ts for rank 20 em ploy ees, 1961-
78 coho rts for ra nk 30 wor ker s, and for 1961-70 coho rts for ran k 40 em ploy ees.
'D um m y term s for en t ry year a re included in each equat ion .
c
Ag e at entry into the rank.
d
Humanit ies major is the omit ted category.
Entry year dummies omit ted from this model .
'Humanit ies and Science major is the omit ted category.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 344/375
3 7
20).
In contrast, college major, which has little effect du ring the trainee years
—possibly because new recruits follow a generalist career path—attains
significance at higher organizational levels, where specialization is more
comm on (Suzuki 1981), with a bu sines s/law major predicting to early
advancement. Thus, what can be said from these data is that trainee
assessments—as indexed by du ration in rank 20—have
profound effect on
advancement prospects over the career course and that, in the higher
corpo rate ranks , college major also predicts to prom otion . As to the
individual-level variables which discriminate among employees during
the trainee period—surely an interesting question—we have no informa
tion. Because of the hom ogeneity of new recruits on most observable
human capital measures and the absence of a work history at the trainee
stage, we lack variables that might differentiate am ong em ployees in this
early career period.
ompensation level
The second dimension in our description of the stratification process
concerns
the
determ inants of mon thly salary. There
is
some consensus that,
at least in past
decades,
seniority has constituted the principal consideration
in
the
calculation of salary level
in
large Japanese firms (Marsh and M annari
1976,
pp. 154-156; Yoshino and Lifson 1986, pp. 152-153). M oreover, w e
have argued that such an arrangem ent meshes well with several distinctive
features of indu strial o rganization in Japan: little differentiation am ong
workers on human capital measures; a practice of rotating employees
among jobs; and a need to mitigate the consequence of failure in the
prom otion com petition. A policy that pegs com pensa tion to seniority
wo uld appear consistent with these organizational practices. At the same
time,
other researchers (e.g., Cole
1971,
pp .
81-84;
Clark 1979, pp. 154-155)
have suggested that Japanese firms have been moving away from a largely
seniority based re ward structure to job specific paym ents and performance
wages.
In a large Japanese firm an em ployee's com pensation level is the su m
of
several com ponen ts: base salary, rank supp lements, family and com muting
allowances, and a bonus paym ent. In the following analysis, we limit our
consideration to the base salary component (which includes additions for
seniority) and the rank supplements.
8
OLS regressions of mo nthly compen
sation (in July 1993) are reported in Table 14.3.
The coefficients in column (1) show the effects of seniority and status
rank on log monthly salary. In this semi-log specification, exponentials of
the rank coefficients can be interpreted as multiplier terms. Thus, holding
seniority constant, the cumulative salary returns to rank—relative to the
base category (rank 30)—are: an 8.5 increase in rank 40 (i.e., a 1.085
multiplier), a
15.8
increase in rank
50,
and a 23.6 increment in rank 60.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 345/375
325
TABLE 14.3 De terminants of Monthly Salary in 1993, OLS Regressions
3
Variable
Constan t
Curren t Rank
b
r
40
50
60
Seniority (years)
(Seniority)
2
Durat ion in Curren t Rank
xlO
2
(years)
Rank 40 x Du rat ion (xlO
2
)
Rank 50 x Durat io n (xlO
2
)
Rank 60 x Dura t ion (xlO
2
)
R
2
=
N =
(1)
2.3580**
.0818**
.1470**
.2121**
.0689**
-.0012**
.962
1723
(2)
12.4114**
.1057**
.2211**
.2926**
.0634**
-.0012**
.6259**
.964
1681
(3)
12.2845**
.0501**
.1384**
.1998**
.0796**
-.0015**
-.3428**
.8258**
1.8912
2.4828**
.967
1681
*p<.05,
**p<.01
Dependent variable is log of monthly base salary plus rank supplem ents. Mean
(yen) = 587,200.
Excluded ca tegory is Rank 30.
In compar ison, holding rank cons tant , the cumulat ive sa lary re turns to
ten ure at the 10,20, and 30 year po ints are, respectively, incre me nts of 77 ,
145%,
and 168% over entry salary.
Co lum ns (2) and (3) repor t m ore complex m odels of the sa lary dete rm i
nati on proce ss . In colu m n (3) term s ha ve been ad de d for du ratio n in rank
an d for in terac t ions betw een this var iable and the rank du m m ies . Because
of the hig h cor re la t ion betwe en d ura t io n and (dura t ion)
2
, a linear specifica
tion of this variable is used.
9
Because the rank effects are now a function of duration, in order to
compare the returns to senior i ty with the returns to s tatus level one must
assess the lat ter at dif ferent d ura tion t imes. At the 20 year po int , the retu rn
to senior i ty
is
a 170% increa se over en try sala ry; at the
30
yea r po int it is little
different: a 182% increment . The
ddition l
salary return to rank, as du rat ion
ran ge s for instan ce from zero to five years , is the follow ing: In ran k 40,
relat ive to the omitte d catego ry (rank
30),
it va rie s from 5.1 for recent ran k
en tran ts to 9.6% at the five ye ar po int. In ran k 50, the rang e is 14.8% to 26.2% ,
and in rank 60 it is 22.1% to 38.3%. W hile these re turn s to s ta tus rank are
ha rdly inco nseq uen tial , they are small in com paris on w ith the increase t ied
to tenure , especially over th e firs t 20 years of em plo ym en t.
To summarize, our results are consis tent with the reports of other
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 346/375
329
investiga tors in several key respects: We find that (a) seniority p lays a
dominant role in compensation determination, especially during the first
two decades of emp loym ent (Pucik 1984b, p . 92); (b) promotion and rank
differentials become important for salary growth after this tenure point
(Yoshino an Lif son
1986, p. 154);
and c) there is evidence of salary compres
sion, in comparison with w estern compensation practices (Pucik 1984a, p .
272;
1974b,
p.
92).
In particular, the most highly paid rank 60 manager in our
data set received 2.3 times the salary of a beginning rank 30 employee;
10
in
the U.S. financial service company the comparable multiplier is approxi
mately 5.0.
The salary practices we have observed in 1993 are not very different from
descriptions pub lished several decades ago
(e.g.,
Ballon 1969a, chap.6;Cole
1971, chap. 3),
despite suggestions of an impending shift to a more western
mode of compensation (e.g., Clark
1979,
p.
155;
Yoshino and Lif son
1986, pp.
152-155).
11
This stability is not surprising to us because the com pensation
strategy of a firm cannot be isolated from other deeply rooted personnel
practices. Rather, it is a key element in a system whose com ponents include
a low variance in hum an capital variables at employm ent
entry,
a generalist
orientation during the training period, an emphasis on solidarity, and
lifetime em ploym ent. Seniority-based compensa tion by a firm is econom i
cally rational in the early years of a coho rt's tenure because there is little
employee differentiation; it is also a reasonable corporate policy in the later
years, after d ifferentiation has taken place, as it limits the consequence of
failure. In a context of lifetime em ploym ent this has to be a critical
consideration since comm itment a nd performance m ust be motivated for
the less successful em ployee .
Formulations of Rank dvancement
What sort of conceptual imagery best describes the attainm ent process in
the Japanese company? Should it be viewed as a) a tournam ent m odel, (b)
a process of cum ulative advantage, c) an example of sponso red m obility,
(d) a gatekeeping operation, or (e) as an instance of contest mobility, in
which the criteria for prom otion to a senior position could be qu ite different
from judgm ents abo ut performance in low ranks? Ou r data set is not
sufficiently rich to distinguish definitively among these alternative expla
nations, thoug h w e are in a position to rule out
some.
Moreover, while the
conceptual imageries may appear distinct when characterized by terse
descriptions, once the explanations are operationalized the predictions
from several of the formulations tend to overlap. In particular, m ode ls (a)
and (b) are difficult to disentangle; the same is true for models (c) and (d).
In its simplest form, the tournament model describes a sequence of
contests in which only winners advance to the next round . According to
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 347/375
330
R ose nb aum (1984, p . 243): This leads to a system in wh ich selections are
cont inual ly occur r ing [ to decide who wil l be promoted and who wil l be]
elim inate d from the tou rna m en t and mo ve d into the category of loser , from
wh ich there is l imi ted op por tu ni ty to adv ance . In Ro senb aum 's
tourna-
ment mod el of careers
w inn ers are opera t ional ized by t ime in rank— they
are the em plo yee s wh o ha ve been pr om ot ed ear ly. A lso, in place of s tr ict
el im ination , the careers m od el posits a s lower rate of pro m otio n for lose rs
an d con sequ ential ly a low er pe ak rank at ret irem ent. There are three
impl ica t ions of the mod el wh ich can be examined with ou r da ta : (a) careers
are s tructured in terms of a sequence of selections, (b) individuals with a
sho rt t ime-d uratio n at one level are l ikely to be pro m ote d ra pid ly from the
nex t rank , an d (c) there is l i tt le op po rtu nity for lose rs to recover . Su pp or t
for the tour nam ent imagery in a la rge Japanese f irm has been repo r ted by
Pucik (1985).
Unfor tunate ly , many of the impl ica t ions of the tournament model a lso
follow from a proce ss of
cumulative advantage
such as w ou ld occur from the
sorting of employees in each corporate rank on the basis of abil i ty or
per forman ce. Even w ithou t a not ion of s t ruc tured com pet i t ion and the
elim inatio n of losers , som e w ork ers will adv anc e rap idly w hile oth ers fall
be hin d. In both formu lations w e sho uld f ind that senior level em plo yee s
progress ive ly pul led ahead of the i r
peers ,
ha vin g spent shor ter dura t ion s in
a rank and consequently having arr ived at each successive s tatus with less
senior i ty. Ro sen bau m (1984, pp . 265-267) at te m pte d to dise nta ng le the tw o
m od els by associa t ing the tourn am ent formula t ion with a label ing pro
cess—rapid pr ior mobi l i ty signals an em ploy ee 's h igh poten t ia l , in co ntras t
wi th r e li ance by man agem ent up on con temp oraneou s eva lua t ions of pe r
formance— but th is emb el l ishm ent do es not he lp to dis t inguish betw een the
formulations with the personnel data available to us .
A sponsorship model (Turner 1960; also see Rosenbaum 1984, p. 17) ,
entails an early selection decis ion an d the assign m ent of em plo ye es to tw o
or m or e tracks. The selection det erm ine s an indiv idu al 's prosp ects of
even tual ly achieving a senior adm inis t ra t ive rank. Co nt ing ent on the t rack
ass ignment , an employee 's subsequent per formance—and his ra te of pro
mo tion f rom m id- level ranks—is re la t ive ly un im po r tant . W hat counts is
the ear ly sor t ing decis ion, which may be followed by special job assign
m ents and gro om ing for e l ite s ta tus . A g tekeeping oper tion sugges ts an
ana log ou s f il ter ing p rocess early in the career course. A subtle d is t inction
betw een the mo dels is tha t spo nso rship is usual ly associa ted with e li te
selection w he rea s ga tek eep ing sugg ests an objective of insu r ing m ini
m u m com petence . Both mo dels , how ever , involve a t racking decision a t an
initial career point.
Turner (1960) also introduced the notion of
contest mobility
in which
decisions ab ou t el ite s tatu s are dela yed w ell into the career cou rse, perm it-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 348/375
33
ting employees an opportunity to overcome poor early evaluations or
otherw ise grow with experience. Ballon's (1969a, pp . 25-26) description
of career dynam ics
in a
large Japanese firm echoes
this
formulation, with his
emphasis on time-in-rank prerequisites for advancement and his stress on
the nuances of socialization
( proper
organizational values ), in place of
narrow w ork performance. The conclusion by Spilerman and Lunde (1991),
from data on the
U.S.
firm,
to
the effect that different talents becom e relevant
to
prom otion decisions as one rises in
the
corporate hierarchy
also
supports
this formulation.
The above alternative models, or com peting imageries, prov ide a frame
work for assessing the structure of the attainment process in the Japanese
firm. In Table 14.4 we examine the effects of duration spells in prio r ranks
on the prom otion rate. Column
(1),
a repeat of Table
14.2,
Panel
B,
column
(1),
is presented for continuity with the earlier analysis and reports a
significant nega tive effect of dura tion in rank 20 (the trainee sta tus) on the
prom otion rate in rank
30.
Because the linear specification of the dura tion
variable might be masking non-linear returns to different interval lengths,
which wo uld be revealing of the consequence of early and late prom o
tions,
we divided the duration variable into four categorical terms: early
prom otion, on-time promotion, late promotion, and very late promotion.
12
On-time , which is defined as the m oda l category, is the reference term .
The results are reported in column (2).
Relative to on-tim e promotion, the tournam ent m odel would predict
rapid advancement from rank
3
for employees wh o were prom oted early
from the prior rank ( winners ), and slow promotion for lagga rds. Ou r
results make clear that laggards do, indeed, have poor advancement
prosp ects, but they fail to show an advantage for early movers. It is also
worth noting that the two laggard categories sum to only
15%
of personnel
(see note 12); in short, rather than a process of gleaning the very best
employees and p reparing them for
elite
positions, the selection mechanism
appears to be oriented to eliminating the chaff—the small proportion of
recruitment errors.
Co lum ns (3) to (5) refer to promotion from rank 40. From column (3) we
observe the effect of duration in rank 30—the prior status—to have no
impact on the prom otion rate. In column (4) we introduce a term for
duration
in
rank
20,
a repetition of
the
model
in
column
(1),
Panel
C
of Table
14.2. As explained in regard to that table, the present formulation is
identical to a model containing variables for seniority and duration in the
prior rank, though for reasons that have been noted the present represen
tation is prefe rred. Even in the presence of this control we find n o effect
of
duration in rank 30. This model does indicate, however, a continued
negative effect of dura tion in the trainee status. In column (5) a mode l is
reported in which the duration terms have been divided into categorical
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 349/375
33
TABLE 14.4 Prom otion Regressions: Effects of Duration in Prior Ranks
8
Variable
A ge
c
(years)
Durat ion in Rank 20:
Cont inuous (years)
Early Promotion
1
Late Promot ion
Very Late Promotion
Durat ion in Rank 30:
Cont inuous (years)
Early Promotion
13
Late Promot ion
Very Late Pro mo t ion
-2LL =
N =
Percent censored =
Promotion
from Rank 3
(1)
.0013
-.1679**
12397.41
1119
11%
(2)
.0171
-.1872
-.2245*
-.6098**
12391.86
1119
11%
(3)
-.0467
.1742
1840.31
300
3 7%
Promotion
from Rank 4
(4)
.1605
-.4770**
-.0567
1830.74
300
3 7%
(5)
.1074
-.1472
-.6813**
-.1540
-.0961
-.0049
1826.05
300
3 7 %
*p<.05, **p<.01
Regressions are Cox proportional hazard models. Clock for the models is
duration in rank. Data cover 1961-82 entry cohorts.
b
Dummy variables for entry year are present in each equation.
c
Age at entry into current rank 30 or 40).
d
Set of four dumm y variables for duration in rank. Deleted term in regressions
is On-Time Promotion, which is defined as the mo de waiting time
11
years for
duration in rank 20; 4 years for dura tion in rank 30).
Because of small N, Very Late Promotion is combined with Late Prom otion
in the the rank 40 regressions.
var iab les to ascer tain whe the r the l inear specifications ha ve ma sk ed dis t inc
t ive non-linear returns to par t icular duration intervals .
13
Again, we fail to
find an effect of duration in the prior grade.
Taken together , these resul ts do not sup po r t a tou rnam ent formula t ion
or a cum ulat ive adv anta ge m od el as a descr ip t ion of the a t ta inme nt process
in the Japa nese com pa ny . There is no evidenc e of sequ ential selections
a m o ng winne r s , a s R os e nba um (1984, p . 243) requ ires for the to ur na m en t
mo del , nor an accumulat ion of rapid p rom otions by the pre sum ably m ore
able em plo yee s. M oreo ver , w e fail to f ind an y indication of a ret urn to
w inn ers , in tha t rapid m obi l ity in the pr ior rank provid es no discernable
advantage over the modal ra te of advancement .
I f our resul ts do not suppor t a tournament model , they are cons is tent
with a selection process in which an ear ly tracking decis ion, based on
evaluat ion s dur in g the tra inee per iod , influences an emp loye e 's a t ta inm ent
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 350/375
333
TABLE 14.5 Distribution of Spell Durations in Rank 30 by Seniority at Entrance
into Rank 30
Seniori ty
at Entrance
to Rank 30
(years)
0
10
11
12
13
14
15
N
1
1
1
3
5
Duration in Rank 30 years?
2 3
80
/
5 27
/ . ,
3
12 110
4
/ '
298
38
6
351
5
14
294
43
3
1
1
356
6
3
89
18
110
7
1
23
10
2
36
8
9
6
15
9 or
m o r e
5
2
1
8
N
27
798
149
20
5
4
1003
Data are for com pleted spells that ended in a promotion.
Mean duration in rank 3 = 4.6 yea rs, median = 5.0 years.
'Seniority at entrance into rank 30 (= mean duration in rank 20) = 11.2 years,
median = 11.0 years .
pro spe cts over the career cou rse. Yet this selection m ech an ism d oes n ot
appear to be one of anoint ing the most promis ing employees , ra ther i t
ope ra tes as a ga tekeep ing process tha t w ee ds ou t the least capable t ra inees .
Co nt ing ent on th is t racking decis ion, the prom otio n ra te in the m idd le rank s
doe s no t ap pe ar consequen t ia l as a de te rmin ant of an em plo yee 's eve ntual
peak s tatus in the f irm.
Th e ne ga tive coefficients of e rly pro mo tion in the pr ior s ta tus , w hich w e
observ e in the d u m m y var iable formula t ions of both the rank 30 and rank
40 regressio ns, d o not reach s ignif icance (colum ns 2 an d 5). N on ethe less , if
the data are organized somewhat differently we can observe that these are
no t cha nce effects, b u t arise from th e very specification of pr om ot io n criteria
in the company.
In Table 14.5 we present a cross- tabulation of duration in rank 30 by
du rat io n in ran k 20 (senior ity at entra nce into rank 30). N ote f irst the
dia go na l in the up p er left corner . It m ak es clear tha t there is a m in im um
req uire m en t of 14 yea rs senior i ty (du ration in rank s 20 pl us 30) before on e
can be cons idered for adv anc em ent to rank 40. As a consequence , emp loy
ees pro m ote d ear ly f rom rank 20 (10 years dura t ion ) mu st w ai t four years
before adv anc em ent , w herea s s low mov ers f rom ra nk 20 can be pro m ote d
again w i thin a year . I t i s th is senior i ty requirem ent w hich pro du ce s th e
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 351/375
334
TABLE 14.6 Distribution of Spell Durations in Rank 40 by Seniority
at Entrance in to Rank 40
a
14
15
Seniority 16
at Entrance
to Rank 40 17
(years)'
18
19
20
N
Duration
in
Rank 4
Years)*
12 or
4 5 6 7 8 9 10 11 mo re
.3
3 1 2 2
1 12 11 6 2 4
3 15 22 12 9 5 8
1
/
6 13 6 5 1 32
/
2 2 2 1 1 1
1 1 1 1 2
1
2 1 12 32 45 32 19 9 16
x
11
36
74
9
5
1
168
a
Data are for completed spells that ended in a promotion.
b
Mean du ration in rank 40
=
8.4 years, median = 8.0 years.
c
Mean seniority at entrance to rank 40 = 16 1 years, med ian = 16.0 years.
ne ga tive effect of ear ly prom otio n in the pr io r grad e, no ted in co lu m ns (2)
an d (5) of Table 14.4, and wh ich un de rm ine s th e possibili ty of a tou rna
m en t /cu m ula t iv e adv antage process desc r ib ing a t t a inmen t in the f irm .
The row s and colum ns in Table 14 5 can be interp reted as dev iation s f rom
their respective me dia ns. In par t icula r , the entr ies in the firs t four co lum ns
represent ear ly promotions f rom grade 30. Thus , among employees who
serv ed 10 years in ran k 20 (row 1), 33 (9/27) w ere prom ot ed ear ly f rom
ran k 30. A m on g em ployee s w i th 11 years du ra t ion in ran k 20, 47 w ere
pro m ote d early f rom grad e 30. For 12 years service the ear ly prom ot ion rate
is also 4 7 ; and for 13 or m ore ye ars in ran k 20 it is 72 . To em ph as iz e tha t
these results are no t idiosyncratic of the par t ic ular rank , w e repo rt in Table
14.6 an analo go us cross- tabulation bet w een sen ior i ty at entra nc e into ran k
40 and dura t ion in r ank 40. T he effects are almo st identical to thos e r ep ort ed
in Table
14.5.
14
To summar ize , these f indings run counter to a content ion tha t rapid
advancement f rom one rank increases an employee 's prospects of ear ly
pro m otio n in the next rank— an ima gery tha t und er l ies both the to urn am ent
m od el and the cumu lat ive adv anta ge process . Ins tead, as a conseq uence of
the f irm's promotion rules , we f ind a tendency for short durations to be
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 352/375
5
coupled w ith long stays
in the
succeeding
rank
Pucik
1985)
concluded that
a tournam ent m ode l adequately fits career evolution in the Japanese com
pany h e studied, how ever w e do not find su ppo rt for this imagery. M ore
over, the specification of a seniority minimum for rank entrance (some
times,
an age minimum) has been reported by other investigators of
Japanese industrial practices (e.g., Yoshino and Lifson 1986, p.147; Marsh
and Mannari 1976,
chap. 7;
also see Koike 1988,
pp.
210-215 for a comparison
between Japan and Eu ropean countries). Thus, even thoug h our assess
m ent is based on data from a single firm, it is unlikely that a tou rnam en t/
cumulative advantage model is widely app licable in the Japanese industrial
context.
Advancement to senior ranks
What more can be said about the attainm ent process? In particular, what
can be ascertained about the selection of senior personnel—the rank 60
managers? Table 14.7 speaks to this question, as well as providing sum
m ary information about career dynamics. These data exam ine curren t
(1993) status for employees hired during the time interval 1962-67—the
entry cohorts that are approaching retirement age. Since we do not have
data on personne l who have left em ployment,
this
samp le provides our best
insight into the achievement of high level pos itions. Table 14.7 describes
current status, contingent on seniority at entrance into lower ranks.
With respect to seniority at entrance into rank 30 (duration in rank 20),
there is a clear division in terms of prospects for reaching status 60 (Panel
A).
On-time entran ts (specified by median seniority) and early entrants are
twice as likely as late arrivals to achieve this level. How ever, there is no
adva ntage to rapid prior advancement— w inners in Rosenbaum 's for
mulation—nor is there a special disadvantage to employees in the very
slowest category. In this specification of
the
seniority categories, some
65%
of employees are in the two groups that show superior prospects for
reaching rank 60.
15
Panel B rep orts the likelihood of reaching the different 1993 ranks,
contingent on seniority at entrance into status
40.
Focusing on status
60,
we
again see evidence of a step function effect: little appa ren t difference in the
attainment prospects of early, on-time, or moderately tardy emp loy
ees, as m easured by prior rate of movem ent, but a clear disadvantage to very
slow movers (15% of the sample), who exhibit an advancement rate into
sta tus 60 that is less than half that of the other categories.
Thus, in ranks 30 and
40,
the advancemen t regime appears to operate as
a gatekeeping process which does not differentiate between rapid and
average
m overs,
nor (in rank
40)
between these group s and m odestly tardy
em ployees, with respect to prospects for achieving elite
status.
A minority
of employees are penalized, namely those with very slow rates of prior
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 353/375
6
TABLE 14.7 Status Rank in
993
by Seniority at Entry into Prior Ranks,
1962-67 Entry C ohort
PANEL A. Rank in
1993,
by Seniority at Entry into Rank 30
Rank in 993
Seniority (mos.) 60 50 40 30 (N)
<132
132
133-144
>144
AH
.64
.61
.32
.30
.51
.18
.27
.26
.20
.26
.18
.11
.38
.40
.21
.00
.01
.04
.10
.02
1.00 (11)
1.00 (148)
1.00
77)
1.00 (10)
1.00 (246)
PANEL B. Rank in 1993, by Seniority at Entry into Rank 40
3
Rank in 993
Seniority (mos.)
<192
192-203
204-215
>215
All
60
.58
.60
.50
.21
.51
50
.31
.27
.17
.24
.25
40
.12
.12
.33
.55
.23
(N)
1.01 (52)
0.99 (81)
1.00 (48)
1.00 (33)
0.99 (214)
PANEL C. Rank in 1993, by Seniority at En try into Rank 50
a
Rank in 993
Seniority (mos.) 60 50
N)
<277
277-288
289-300
301-312
>312
.96
.82
.70
.47
.19
.04
.18
.30
.53
.81
1.00
1.00
1.00
1.00
1.00
(26)
(39)
(30)
(19)
(26)
All
.66 .34
1.00 (140)
Entries are percentages of the seniority group .
mobili ty, yet , even here, there is substantial opportunity for recovery,
wh ich is no t a feature of the to urn am en t imag ery (R osen bau m 1984, p.42).
Som e 20 - 30 perc ent of the lag ga rd catego ry do succeed in reach ing the
highes t s ta tus .
Pro m otio n from sta tus 50 exhibits a different patt ern. W here as, in low er
ranks ,
the advan cem ent regim e appe ars to opera te by cur ta i l ing the a t ta in
m ents of a min or i ty of em ployee s (a ga teke eping o pera t ion) , we now see
evide nce of progressive differentiation in term s of pri or service (or a ge
16
). In
par t icular , em ployee s wh o hav e advan ced the mo st rapidly to rank 50 ha ve
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 354/375
7
TABLE 14.8 Status Rank in
1993
by Seniority at Entry into Prior Ranks, 1962-67
Entry Cohort, Employees with 132 or Fewer M onths in Grade
20
Only
PANEL A. Rank in
1993,
by Seniority at E ntry into R ank 30
a
Rank in 993
Seniority
(mos.)
60
50 or
lower
PANEL
B.
Rank in
1993,
by Seniority at Entry into R ank
40
Rank in 993
PANEL C Rank in 1993, by Seniority at En try into Rank 50
Rank
in
993
N)
<132
132
All
.64
.61
.61
.36
.39
.39
1.00
1.00
1.00
(11)
(148)
(159)
Seniority
(mos.)
<192
192-203
204-215
>215
All
60
.62
.67
.58
.38
.62
50 or
lower
.38
.33
.42
.62
.38
1.00
1.00
1.00
1.00
1.00
(N)
(39)
(66)
(24)
(8)
(137)
Seniority (mos.)
<277
277-288
289-300
301-312
>312
All
60
.96
.81
.65
.54
.17
.70
50
.04
.19
.35
.46
.83
.30
1.00
1.00
1.00
1.00
1.00
1.00
(N)
(23)
(31)
(23)
(13)
(12)
(102)
Entries are percen tages of the seniority g rou p.
the best prosp ects of reachin g ran k 60; the s lower the pr io r progre ssion , the
poorer an individual ' s chances .
A more ref ined assessment of the a t ta inment process can be obta ined
from Table 14.8, in wh ich the calcula tions of Table
14 7
hav e been r epea ted
bu t w i th the lag ga rds in Panel A rem ov ed (employees in the las t two
rows) . W e no w obs erve a vir tual absen ce of a tenu re effect at entran ce into
rank 40 on one 's prospects of achieving high s ta tus in the organiza t ion
(Panel
B);
in sho rt, little differentiation takes place at this caree r sta ge . Th e
fall off w ith te nu re , no ted in Pan el B of Tab le
14.7,
w as larg ely a n artifact of
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 355/375
338
the inclusion in that table of employees who failed the gate-keeping cut.
(The fall-off with tenure that remains in Panel
refers to
5%
of em ployees
and the N is too small to establish statistically that the rate is below the
category average.) With respect to the impact of seniority a t entrance into
rank
50 ,
our results are very different from the lower grades and exh ibit a
strong tenure effect (Panel C).
In regard to conceptual imagery, these findings add up to what can be
considered a two step process First, a screening or gate-keeping decision is
m ade in the lowest grades
to
weed out poor performers, as indexed by very
slow rates of prior mobility. The remaining employees— perhaps
70%
of an
entry cohort—are no t differentiated with respect to prospects for attaining
elite status as they move through the middle ranks of the organization.
Possibly, this is intended to maintain m orale by avo iding an early labeling
of employees as winners and losers —an understandable strategy in a
firm in which demoralized workers cannot be dismissed. Possibly, as
Ballon (1969a, pp . 25-26) has con tended, elite selection is delayed because
the ability to perform senior tasks effectively is revealed late in the career
course. At any rate, it is only at the second stage, prom otion from rank 50,
where there is clear evidence of a graduated return to prior mobility and
where length of service
is an
effective indicator of prospects for atta ining the
highest status.
onclusions
In contrast with Pucik's (1984b; 1985) assessment of the attainment
process, we find little evidence to support the imagery of a tournament
m odel of careers in the Japanese financial service com pany . As we have
noted, in
a
context of lifetime em ployment
it
makes sense for
a
firm
to
delay
revealing (and perh aps deciding) who will be permitted
to
reach the highest
ranks.
Other elements of the reward structure—such as seniority based
com pensa tion, with only small additions to salary for achieved rank—also
contribute to cohesion and to the morale of the w orkforce.
There have been suggestions of impend ing change in the structure of
careers and rewards in Japanese industry, such as hiring experienced
workers and increasing salary paym ents for achieved rank
(e.g.,
Cole
1971,
chaps. 3 4;
Marsh and M annari
1976, pp. 307-314).
H owever, in the firm we
have studied there
is
little evidence to suggest m uch movem ent away from
the kind of reward structure that was described some 25 years ago (e.g.,
Ballon 1969b). Moreover, because of the tight interrelation of the compo
nent features of the reward structure, we suspect that when change do es
come it will not be g radual or piecemeal.
Finally, what can be said about w ho achieves elite status? Actually very
little.
There is markedly little differentiation am ong emp loyees through
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 356/375
9
rank 40, in pa r t beca use the entry cohor ts are ho m og ene ou s w i th respect to
ed uc atio n, age, ge nd er , and ( lack of) pr ior experience. In the later career
stages, where differentiat ion does take place, i t appears to ref lect subtle
con side ration s of suitab il i ty an d va lues (Ballon 1969a, ch ap ter 1) in
add i t ion to job per form ance— me asures tha t a re not usu al ly presen t in the
pe rso nn el records m ad e available to researchers . Interestingly, there is also
ev iden ce from th e U.S. f inancial service co m pa ny (Sp ilerman and L un de
1991) that the role of education in promotion decis ions is weaker in high
co rpo rate ran ks tha n in the mid dle levels . There, too, i t w as arg ue d tha t the
criteria for effective p erfo rm an ce are quite different for elite m an ag er s th an
for mid -level em plo ye es, that consid eration s of social s tyle an d perso nali ty
play a greater role in the senior rank s than do assessm ents of cognitive ability.
Notes
This research was supported by a grant from the U.S.-Israel Binational
Foundation (to Spilerman) and by an Abe Fellowship from the Social Science
Research Council (to Ishida). We would like to thank Kuo-Hsien Su for assistance
with the computations.
1.
In
Japan,
size of firm
is a
key differentiating variable
in
the stratification system .
Both the status characteristics of workers (educational attainment, household
wealth) and the rewa rds of employm ent (occupational prestige, income) are more
favorable for employees of large companies (Ishida 1993, p p . 208-226).
2. Koike
(1991),
in
one
of
the
few studies which addressed the issue of specializa
tion and skill formation in Japanese companies, documents the practice of special
ization am ong w hite-collar workers bu t also shows tha t employees tend to experi
ence a broad range of jobs within a specialty.
3. We note that there is a counterpart literature with respect to attainment in
western firms which argues that the personality demands of jobs at different
organ izational levels can be quite distinct Moreover,
the
character type and coping
style of an em ployee that is effective in low organizational ranks may be dysfunc
tional in higher g rade s (Silver and Spilerman 1991).
4. Since enactment of the Equal Employment O pportunity Law in 1986 wom en
have m oved into the m anagerial ranks in Japan, although they are still very few in
number. Only one percent of career-track employees in the financial service
company are women.
5. We lack data on emp loym ent history prior to entry in to the U.S. firm. This
estimate is based on the proportion of hires who are age 5 and older.
6 The Japanese company uses the term non-management emp loyees and these
wo rkers are placed in the lifetime queue from the time of employment. We prefer
the term traine es because this is the better fitting description in the American
context. These workers are not promoted from lesser positions based on perfor
mance but are hired directly into this fixed-duration status from which most
progress to management positions.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 357/375
MO
7 I t is m ore com m on to in t roduce sen ior i ty and dura t ion in pr ior rank as the
tem po ral variables. In a m ode l w ith exact ly tw o prior ran ks, as in the Panel C
regressions, this specificat ion can be obtained by a simple t ransformation of the
rep orte d covariates, s ince seniori ty = dur at ion in rank 20 + du rat io n in rank
30 . In particular, the model equivalent to column (1) has coefficients for seniority
an d d ura t ion in rank 30 equal to
-.0397**
an d .0350**, respe ctively . The ne ga tive
effect of seniori ty on the prom otion rate mak es sense in term s of the repo rts of oth er
resear chers (e .g. , Ro sen bau m 1984, pp . 169-170); the po sitive effect of dur at io n in the
prior grade is less comprehensible unt i l one observes that , with seniori ty held
cons tant , a posi t ive dur at io n in rank 30 is equiv alent to a nega t ive du rat ion in rank
20— which is the effect we report . O ur prese ntat ion of the tw o dura t ion t erm s as
covariates, in place of seniori ty and durat ion in prior rank, permits a more direct
interpretat ion of the findings.
8. Ente rtainm ent al lowances are om it ted from this invest igat ion. Since the
a l lowances increase wi th rank , our resu l t s somewhat underes t imate the t rue rank
effects on monthly salary.
9. De spi te the sma ll increase in R
in m oving from equat ion 2) to equat ion (3), the
set of ad de d term s is highly significant base d o n a conve ntional F-test.
10 .
This valu e is calculated from th e m onth ly salary data and is so m ew ha t sm aller
than the regress ion-based mu l t ip lie rs reported in the preced ing par agr aph . No te ,
also ,
that wh ile Pucik (1984a, 1984b) appl ies the term salary com pre ssio n to the
small salary dispersion found when seniori ty is held constant , our comparison
refers to unadjusted salary figures.
11. Part of the reason for the stabil i ty in com pen sat ion pract ices derives from the
fact tha t the Jap ane se com pa ny is in the financial service sector. Th e seniority
com pon ent o f com pensa t ion h as been t rad i t iona lly h igh in th i s indu st ry (Higuehi
1991).
12 .
The dur at ion in rank 20 term s are defined as fol lows: early pr om oti on (10
yea rs in rank) contains 18% of obse rvat ion s, on-t im e pr om oti on (11 years)
contain s 66% of obse rvat ions , la te pro m oti on (12 years) contains 12% of obser va
t ions, and very late pr om oti on (13 or m ore years in rank) conta ins 3 % of
obse rvat io ns. Th ere w ere no pro m otion s before 10 years in rank.
13.
The dura t ion in rank 30 categories are defined as follows: early pr om oti on
(1-3 years in rank) contains 14% of obse rvat ion s, on-t im e pr om oti on (4 years)
con tains 36% of obse rvat ion s, la te pro m otio n (5 years) contains 35% of obser va
t ions, an d very late pr om otio n (6 or m or e years in rank) contains 16% of
observa t ions . O n- t im e prom ot ion , which is def ined as the m odal ca tegory , is the
reference term in the regressions.
14. H ow m ateria l is the exclusion of inco m plete spells from Tab les 14.5 an d 14.6?
This has no conseq uence w ith respect to the d iagonal pat tern in the up pe r left corner
of the tables because the diagonals derive from short durat ions in the adjacent
gra de s. The exclusion of long spel ls is potent ial ly m ore seriou s in the calculat ion of
ear ly prom ot ion
rates,
bu t an exam inat ion of incom plete spel ls (dura t ions in curr ent
state) reveals no major dispari t ies from the resul ts we have presented.
15.
Th ere is m uc h lum pin es s in the data, w hic h pe rm its li t t le flexibility in the
assig nm ent to seniori ty categories. In part icular, 148 em ploy ees w ere p rom ote d
exact ly
11
ye ars after entranc e into the com pan y. Similarly , abo ut
95%
of em ploy ees
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 358/375
341
in the next category we re prom oted at exactly the 144 month po int. Thus, there is
no possibility for exploring the sensitivity of the reported findings to the seniority
category bou nda ries. Note, also, that no employee was prom oted to rank 30 with
less than 10 years service.
16.
Because of the small age variance in each entry cohort, the argument of this
section, which stresses seniority, can be formulated equivalently in terms of em
ployee age.
eferences
Ballon, Robert
J
1969a. The Japanese Dim ensions of Indus trial Ente rprise. Pp. 3-
40 in Robert J Ballon (ed.),
The Japanese Em ployee.
Tokyo: Sophia U niversity.
. 1969b. Lifelong Rem uneration System . Pp. 123-166 in Robert J. Ballon
(ed.),
The Japanese
Em ployee. Tokyo: Sophia University.
Blossfeld, Hans-Peter, Hamerle, Alfred, and Karl U. Mayer. 1989.
Event History
Analysis. H illsdale: Erlbaum Associates.
Clark, Rodney. 1979.
The Japanese Company.
New Haven: Yale University Press.
Cole, Robert. E. 1971. Japanese Blue
Collar:
The Changing Tradition. Berkeley: Univer
sity of California Press.
. 1979. Work
Mobility and Participation.
Berkeley: University of California
Press.
Dore, Ronald.
1973.
British Factory-Japanese Factory. Berkeley: University of Califor
nia Press.
Funahashi, Naomichi. 1973. The Industrial Reward System: Wages and Benefits.
Chap.
10 in
Workers and Employers in Japan
K Okochi,
B.
K arsh, and S B Levine
(eds.). Princeton: University of Princeton P ress.
Foy, Nancy. 1975. The Sun Never Sets on IBM. New York: William M orrow.
Hirono, Ryokichi. 1969. Personnel Managem ent in Foreign Corporations. Pp.
251-272
in
The Japanese
Em ployee Robert
J
Ballon (ed.). Tokyo: Sophia University.
Ishida, Hiroshi. 1993.
Social Mobility in Contemporary Japan.
Stanford: Stanford
University Press.
Koike, Kazuo. 1988.
Understanding Industrial Relations in Modern Japan.
London:
MacMillan P ress.
Marsh, Robert
M.
and Hiroshi
M annari. 1976. Modernization and the Japanese Factory.
Princeton: Princeton University Press.
OECD (Organization for Economic Cooperation and Development). 1973. Man-
power Policy in Japan.
Paris: Organization for Economic Cooperation and Devel
opment.
Okochi, Kazuo; Karsh, Bernard; and Solomon B. Levine. 1974. The Japanese
Industrial Relations System:
A
Summ ary. Pp. 485-517 in
Workers and Employers
in Japan
K. Okochi, B. Karsh, and S.B. Levine (eds.). Princeton: University of
Princeton Press.
Peck, Merton and Shuji Tam ura. 1976. Technology . In
Asia s New
Giant Hugh
Patrick and Henry Rosovsky (eds.). Washington, D .C.: Brookings Institution.
Petersen, Trond and Seymour Spilerman . 1990. Job Qu its from an Internal Labor
Market. Pp.
69-95
in Event H istory Analysis in Life Course Research Karl U. Mayer
and Nancy B. Tum a (eds.). Madison: University of W isconsin Press.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 359/375
34
Pucik, Vladimir. 1984a. White-Collar Hum an Resource Managem ent in Large
Japanese Manufacturing
Firms. Human
Resource
Management
3 (3), pp. 257-276.
. 1984b. White Collar Human Resource Managem ent: A Comparison of the
U.S.
and Japanese Automobile Industries.
Columbia Journal of World Business
19
(3), pp . 87-94.
. 1985. Promotion Patterns in a Japanese Com pany. Columbia Journal of
World Business
20 (3), pp. 73-79.
Rohlen, Thom as.
1974. For Harmony and Strength
Berkeley: University of California
Press.
Roland Alan. 1988. In Search o f Self in India and Japan Princeton: Princeton University
Press.
Rosenbaum, James E. 1984. Career
Mobility
in a
Corporate Hierarchy
New York:
Academ ic Press.
Rosenbaum, James and Takehiko
Kariya. 1989.
From High School to
Work:
M arket
and Institutional Mechanisms in Japan.
American Journal of Sociology
94 (6), pp.
1334-1365.
Sibson, Robert E. 1981. Com pensation New York: AMACOM.
Silver, Catherine an d Seymour Spilerman.
1991.
Psychoanalytic Perspectives on
Occupational Choice and Attainment.
Research in Social Stratification and
Mobil-
ity 9 pp. 181-214.
Spilerman , Seymour and Tormod Lunde. 1991. Features of Educa tional Attain
ment and Job Promotion Prospects.
American Journal of Sociology 9
(3), pp . 689-
720.
and Trond Petersen.
1994.
Occupational Structure, Determinants of Promo
tion, and Gender Differences in Attainm ent. Subm itted for pub lication.
Suzuki, N. 1981.
Management and the Industrial Structure in Japan
New York:
Pergamon Press.
Turner, Ralph. 1960. Modes of Social Ascent Through Education: Sponsored and
Contest Mobility.
American Sociological Review
25, pp . 855-867.
Wallace, Marc J. and C harles H. Fay 1983.
Compensaton Theory and Practice
Belmont,
Calif.: Kent Publishing Co.
Yoshino, M.Y. and Thomas
B.
Lifson. 1986.
The Invisible
Link.
Cam bridge, M ass.:
MIT Press.
Wolferen, Karl van and R. Taggart M urphy.
1994.
Tough Love for Japan.
New York
Times February 20,1994, p. IV-13.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 360/375
5
C hang ing Contexts of Careers:
Trends in Labor M arket S tructures
an d Some Implications for
Labor Force O utcom es
Arne
L
Kalleberg
University of North
arolina
at
hapel
Hill
Labor market structures are major institutional contexts which help to
define the structures and processes that generate earnings em ploym ent
contracts patterns of job changing and unemploy m ent and other aspects of
careers. In the past quarter century considerable progress has been m ade
in und erstan ding both the structure of labor markets and their impacts on
these important labor force outcom es. Con ceptions of labor marke ts have
not been static; the view do minating at a particular time reflects in par t
theoretical dev elop ments in sociology and econom ics as well as political
social and econom ic events that heighten concerns with particular issues
and problem s and rend er other questions moot. The dom inant conception
of labor markets at a par ticular time in turn affects the kinds of labor force
issues and aspects of careers that are deem ed impo rtant to study .
This paper discusses some of the main conceptions of labor market
structures in the past quarter century and how they have influenced
research and thinking about careers. The first section sum marizes three
approaches to labor markets: dua lism; segmentation; and flexibility. then
speculate on some implications of current changes for future labor ma rke t
structures and various career-related o utcomes.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 361/375
344
pproa ches to Labor M arkets
Dualism
Th e resurgen ce of interest in labor mar kets in the la te 1960s w as p ro m pt ed
by concerns am ong radica l econom is ts to account for w hy the b lack u rb an
poor in the United States remained poor despite efforts to al leviate their
po ver ty (see the sum m ary in Go rdon , 1972). The answer w as beca use the
po or part ic ip ated in the secon dary labor m arke t , the segm ent of the eco nom y
cha racte rized by low w age s, job instabil i ty, an d gene ral ly bad jobs. The
si tuatio n in the seco nda ry labor m ark ets wa s radical ly different from that in
the pr imary market , which cons is ted of s table , wel l pa id , and genera l ly
des i rable em ploy m ent . Th oug h the im agery of two dis t inc t labor m arket
segments was important , the exis tence of
tw o
l abor markets was less
important than the idea that there was a radical discontinuity in labor
market processes between them (see Berger and Piore, 1980).
Dual labor market theor is ts a rgued tha t the or ig ins of pr imary and
secondary markets were rooted in workers charac ter is t ics (such as the i r
beh avio ral preferences for s table vs . uns table w ork ), as we ll as in tech nol
og y an d skill (for a rev iew , see K allebe rg and S 0re nse n, 1979). S om e of these
m ech an ism s (such as technolog y) are consis tent with neoclassical econo m ic
theory; o thers (such as the assu m pt ion tha t peo ples preferences a nd
at t i tud es a ren t exog enous to the labor market ) a re not . The dua l labor
market theory was fairly static in its focus on the existence of unequal
s t ruc tu res of pay a nd s tabil i ty , desp i te it s assu m pt io ns ab ou t the polar iza
t ion of jobs acco m pan yin g the evolutio n of capital ism.
The idea of labor market dua l ism soon expanded beyond ghet to labor
markets to inc lude the ana lys is of na t iona l labor markets in advanced
industria l societ ies , and the micro-level dual labor market theory was
linked to the m acro-level du al econ om y theory (e .g. , Av eri t t , 1968; Blue-
s tone , 1970). The lat ter focused o n the du ali ty be tw een co re an d pe rip he ry
indus t r ies and f i rms , and provided both economis ts and soc io logis ts wi th
a useful way of conceptualizing quali ta t ive divis ions result ing from the
evo lution of the orga niza tion of capital ist pro du ctio n. Tak en togethe r, these
dual theor ies pos i ted a correspondence be tween pr imary labor market
segm ent jobs and core sec tor indus t r ies and f i rms , and be tw een second ary
labor mar ket segm ent jobs an d per iph ery sec tor f i rms and indu s t r ies .
Dualism and Research on Careers
The firs t generat ion of dualis t
research was mainly concerned wi th demons tra t ing tha t there were radi
cally different pro cesses of earn ings de term inatio n in the tw o segm ents . I t
a lso sou gh t to assess the extent of inter-se gm ent m obil i ty; the absence of
mobi l i ty be tween segments was an assumpt ion in tegra l to the v iew tha t
there was soc ia l c losure wi th in markets . Studies supported the v iew tha t
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 362/375
45
there were few mobili ty channels that people might use to exit f rom the
seco nda ry mark et . This focused a t tent ion on the cons tra ints faced by po or
w orke rs w ho soug ht to exper ience up w ar d mobi l i ty by inves t ing in edu ca
t ion an d other forms of h u m a n capital . Trainin g and skil ls w ere no t
par t icular ly va l ued in secondary labor ma rket jobs , no r useful as reso urces
for upward mobi l i ty .
Dual labor market researchers a lso examined how and why cer ta in
categor ies of workers—usual ly whi te males in the Uni ted S ta tes—were
integ rated in to the core of the socioeconom ic system w heth er vie w ed in
terms of core sec tors or pr imary labor markets ) , whi le others—women,
blacks , an d o ther mino r i t ies—w ere ass igned to the per iphe ry of th is sys tem
by lack of human capi ta l , d iscr iminat ion, or inappropr ia te behaviora l
dispos i t ion s . These s tudies he lped to expla in w hy these minor i t ies found
them selves in a pe rm an en t precar ious socioeconomic posit ion see Kalleberg
an d S 0ren sen, 1979, for a review of this research) .
Segmentation
Interest in labor markets accelerated rapidly in the 1970s. Sociologists
w ho s ou gh t to exp lain patt ern s of social s trat ification saw labo r m ark ets as
a w ay of inc orp oratin g in sti tutional an d s tructu ral var iables in the mo de l of
the socioeconomic achievem ent process tha t or iginated in Blau an d D un can
1967),
and in explan at ions of mobi l i ty and inco me inequal i ty mo re gener
ally. Radical eco nom ists an d M arxists found the no tion of labor m ark ets
attractive for exp lainin g chang es in the orga nizatio n of w ork an d the labo r
proc ess that w ere associated with capital is t de vel op m en t. This seco nd
genera t ion of labor market researchers recognized tha t dua l labor m arke t
theory overs impl i f ied the complex mec han isms tha t c rea te an d m ainta in
labor market s tructures in industr ial societies . They thus focused more on
the divers i ty and forms of labor m arket segm enta t ion.
1
There were tw o m ain thru s ts of research on labor m arke t segm enta t ion
in th e 1970s a n d early 1980s. Th e first, aggregated approach sought to
descr ibe the s tructure of labor markets in a society as a whole by, for
exam ple , der iving c lus ters of re lat ive ly ho m og en ou s occu pat ion - indus try
catego ries e .g., Fre ed m an , 1976; see also Spile rma n, 1977). O the r re search
ers used t ime ser ies data to analyze changes in inequali t ies between
core -pe r iphery indus t ry g rou ps an d /o r be tw een pr imary- second ary occu
pa tio n gro up s. Reich 1984), for exa m ple, arg ue d that the pos t-W orld W ar
I I Uni ted S ta tes was marked by a segmenta t ion phase character ized by
growing dis t inc t ions among both occupat ions and indus tr ies .
A second,
disaggregated
app roach to segm enta t ion research focused on
specif ic occupat ion s and orga niza t ions . This ap pro ach centered on the
conc ept of the in ternal labor m arke t ILM), wh ich wa s original ly discusse d
by insti tution al econ om ists in the 1950s. ILMs are job lad de rs cha racter ized
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 363/375
346
by ent ry a t the bot tom and upward movement assoc ia ted wi th a progres
s ive dev elop m ent of sk il l and k now ledge ; they are found in bo th occup a
tions (OILMs) and firms (FILMs) (see Althauser and Kalleberg, 1981).
Both aggregated and d isaggregated approaches sought to account for
the processes by which in te rna l iza t ion occurred wi th in f i rm and occupa
t iona l labor m arkets . The h ierarchica l— as opp osed to ma rket— gov ernan ce
of employment relat ions was argued to result from the operat ion of social
forces (such as employer control , unionization, l icensing, educational cre
dentials) as well as from technical factors (such as skill specificity and the
imperatives of on-the-job training).
Som e wri te rs recognized explic it ly tha t labor m arke t segm enta t io n do es
no t procee d in a s teady , l inear fashion, bu t rath er reflects bro ad er econ om ic
cond i t ions . Org aniza t ions seek to ada pt to bo om and bu s t bus ines s cycles
by incre asing or decr easing their em plo ym en t. Sen genb erger (1981), for
exam ple , show ed tha t Germ an f irms decreased the ir nu m ber of pr ima ry- typ e
jobs duriiig the mid-1970s recession, thereby crea ting a smaller core wo rkforce
and increas ing the number of shor t - te rm contrac ts and employee leas ing
arra ng em ents . This s t ra tegy represen ted a re t renchm ent f rom em plo ye rs
emphasis on internal labor markets to bind scarce labor to the f irm in the
prev iou s decade , wh en unem ploy m ent w as ve ry low. Th i s t endency i s
consis tent with a common response of employers to crises , which is to
marginalize or externalize part of their workforces (see below).
Segmen tation and Research on Careers
The notion of internal labor
markets he lped researchers expla in how labor market s t ruc tures shaped
ind ivid ua ls careers . The concept of FILMs pro ve d par t icular ly useful for
s tud yin g d if fe rences in career pa t te rn s and pro m ot ion o ppo rtuni t ie s wi th in
organiza t ions , and job ladders provided a way to account for pa t te rns of
m ob ility w ith in firms (e.g., Baron et al., 1986; C oh en an d Pfeffer, 1986;
Pfeffer an d Co hen , 1984; Sp ilerm an and Ish ida, this vo lum e; see the r eview
in A lthau ser, 1989). FILMs w ere also central to descr ipt ion s of b ure auc rat ic
con trol system s with in large f irms in the post-W orld W ar II per iod . Th us,
FILMs enab led m anage rs to con tro l work e rs by p rom is ing them op po r tu
nit ies for future ad va nce m en t in retu rn for their com m itm ent a nd loyalty to
the organiza t ion (Ed wa rds , 1979) .
Flexibility
Research on labor market segmentat ion during the 1960s and 1970s—
especial ly in the United S tates , bu t a lso in m an y oth er ind ustria l ized
nat ion s— wa s cond ucted wi th in a context of re la tive s tabil ity an d grow th.
Du ring pe r iod s of rela t ive prosper i ty , i ssues such as career adv anc em ent
and earnings d i f fe rences among fu l l - t ime workers become important to
s tudy. Accordingly , ques t ions of labor market uncer ta in ty and s t ruc tura l
cha ng e we re given relat ively min or roles (see Ro senberg, 1989). y con trast ,
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 364/375
347
the 1980s were marked by diminished growth and high unemployment
rates in m any industrialized
nations.
Product markets became increasingly
complex, unce rtain, variable, and m arked by rapid change (Piore and Sabel,
1984).
These altered economic conditions prom pted a shift in labor marke t
research away from topics of segm entation, and tow ard a focus on dyn am ic
changes in labor market structures and how they could be restructured to
increase flexibility (Rosenberg, 1989).
The issue of labor market flexibility stimulated considerable debate on
both sides of the Atlantic. It was generally agreed that em ployers need the
flexibility to reorganize em ploym ent relations
in
response to rapid develop
ments in technology, growing international competition in produ ct mar
kets,
and changes in labor markets such as the increased labor force
participation of wom en.
The flexibility deba te focused on the firm, and a ttention cen tered on three
main types of
flexi ility
(Atkinson,
1987;
Boyer,
1987; Lane , 1989;
Meulders
and Wilkin, 1987; Piore, 1986; Rosenberg, 1989):
{1 functional lexibility
he
extent to which employers are able to move their existing (relatively
permanent, full-time) employees from one task or department to another,
a n d /o r to reorgan ize the content of jobs to adjust to changing technologies
and economic conditions; 2)
numeric l or employment
lexibility he employ
ers '
freedom to alter workers' ho urs (numerical) or to adjust the sizes of their
workforces (em ployment) to respond to changing conditions; and 3)
w ge
flexibility,
the ability of employers to adjust w ages downw ard when eco
nomic conditions warranted.
Maximizing all three types of flexibility simultaneously is difficult:
strategies designed to achieve functional flexibility may undermine the
ability of emp loyers to obtain num erical a nd /o r wage flexibility, since the
conditions that promote the former (e.g., commitment to the firm) are
incompatible with policies that emphasize layoffs and wage reductions.
One way that an organ ization m ight balance its needs for various types of
flexibility is suggested by the notion of the flexible firm see Atkinson, 1987,
Pollert, 1988; also called the core-periphery model— Osterman,
1988;
or
the sham rock organization— Handy , 1989) (see Figure 15.1). This idea
applies dualist no tions of core-periphery to the firm level. It suggests that
there are divergent processes of labor market segmentation w ithin the firm
that are based on a combination of market and hierarchical governance
system s. Optimal cost effectiveness is achieved by d ividing the labor force
into fixed and variable com ponents. The fixed com ponent consists of a core
of highly com mitted, functionally flexible w orkers wh o are integra ted into
the firm via long-term employment contracts, are trained to perform
multiple skills, and are often em powered to participate in decision-making.
This is similar to the salaried internal labor marke t mode l that combines
relatively flexible administrative procedures with high emp loymen t secu-
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 365/375
48
FIGURE
15 1
The Flexible Firm
1
1
Core (Full-time, Long-term)
I Periphery (Part-time)
H i l l Periphery (Temporary)
, Hill Externalized (Sub-contracting)
rity (thu s achie ving hi gh func tional flexibility at the ex pen se of low em pl oy
m en t flexibility).
2
Two addi t ional , var iable components provide organiza
t ions wi th wage and numer ica l /employment f lexibi l i ty : (1) a per iphery
group of temporary and par t - t ime workers who are only weakly a t tached
to the org an izat ion ; an d (2) an exte rnaliz ed set of activities (s ub co ntra ctin g)
that a re per formed by people who are not permanent organiza t ional
m em be rs (see A tkinso n, 1987; H an dy , 1989; Atch ison, 1991).
3
A co un try 's inst i tutional , poli t ical-econom ic envir on m en t will affect the
extent to which employers are able to segment their workforces in ways
su gg es te d by the m od el of the flexible firm. In the Unite d States, the relativ e
w eak ness of union s and em ploy m ent a t wi l l enable em ploy ers to achieve
re la t ive ly high wage and numer ica l /employment f lexibi l i ty .
4
By con trast ,
in Germany, the combinat ion of unions ' s t rength, dominance of occupa
t ional markets and s t rong sys tems of vocat ional t ra ining has cons tra ined
the de ve lop m en t of pe r iph eral labor forces w ithin f irms. This ha s m oti
vated G erm an co mp anies to place grea ter em pha s is on fos ter ing funct ional
f lexibil ity in order to ad ap t to chan ges in bu sine ss condit ion s and techn ol
ogy (L ane, 1989; see also the stu die s in Boyer, 1988; Se ng en be rge r an d
Koh ler , 1987) . The Japan ese form of hu m an resourc es m an ag em en t is of ten
cited as th e ex em pla r of the f lexible f irm. I t ap pe ars to ach ieve bo th
functional
and
numer ica l /employment f lexibi l i ty by: (1) segment ing the
orga niza t io n 's labor force in to a core se t of pe rm an en t job s /e m plo ye es
honko and a pe r iph ery grou p of t empora ry , con t ingen t job s /em plo yee s
(Koike, 1978); an d (2) a w ell-d ev elo pe d sys tem of sub con tracti ng .
5
The abil i ty and will ingness of organizations in par t icular nations to
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 366/375
349
seg m en t their labor forces also d ep en d on their s ize. Large f irms in the
Un ited S ta tes tend to re ly mo re heavi ly on temp orary w orke rs than do smal l
and m ed ium -s ize d f i rms , because larger organ iza t ions have a hard er t ime
fir ing people and adjusting their hir ing practices to changing condit ions
( M a n g u m e t al., 1985). In Jap an , on the othe r h an d , the ratio of reg ula r (core
or long- term) work ers to tem po rary an d d ay laborers decreases as the s ize
of the f irm increases: temporary workers are more commonly found in
smal l com pan ies . Larger Japanese organiza t ion s rely m ore on var iou s kin ds
of subcontrac t ing re la t ionships , as opposed to hi r ing large numbers of
tempora ry worker s .
Ind ustr i al differences will also affect th e de gree of labor ma rke t seg m en
tation w ithin f irms. The use of tem po rar ies in the U.S. is m ore co m m on in
indus t r ie s exposed to f luc tuat ing p roduc t dem an d (M angum e t al., 1985). In
Japan , the na tur e of subcontrac t ing var ies amon g indus tr ies : in ind us tr ies
such as s teel , shi pb uild ing or chem icals , there is w ide sp rea d use of
sh g iko
or employees of other companies who are leased to the parent f irm on a
tem po rary bas is ; in the autom obi le , e lec t rica l mac hinery a nd cons truct ion
indus tr ies , ac tiv i ties are subcontrac ted out to smal ler fi rms w ho se em ploy
ees do not actually work in the parent f irm (Tokunaga, 1984) . In Norway,
highly unionized banks have re la t ive ly large core and smal l per iphera l
labor forces; while the opposite is true in the relat ively weakly unionized
retail industry.
Flexibility and Research on Careers Concerns about flexibility gave
ad de d s ignificance and impe tus to research on un em plo ym en t and pat te rns
of labor force a t tachm ent . U ne m plo ym ent ra tes w ere found to ha ve
increased less in the late 1970s and early 1980s in countries where wage
flexibil ity w as hig h (U.S., Japan) th an in Eu rop ean natio ns th at had less
w ag e flexibility (Boy er, 1987). Th is is con sistent w ith th e assu m pt io n of
economis ts ' market models tha t wage and employment f lexibi l i ty are
trade-offs : increasing wag es redu ces em plo ym en t; an d vice versa (see
Piore, 1986).
Research on temporary, par t - t ime, and subcontrac t ing employment
relat ions also bec am e m ore preva lent and p ressin g in the
1980s,
as long -term
em plo ym en t contrac ts becam e less com mo n. The term cont ing ent em
plo ym en t w as coined in the mid-1980s to descr ibe the gro w ing t rend
tow ard these types of non - t radi t ional em plo ym ent re la tions tha t hav e
em erg ed from the conce rns for wo rkpla ce an d emp lo ym en t f lexibil ity.
S tudies found tha t the growth of cont ingent employment in the Uni ted
States was due in par t to the growing labor force par t icipation of women
(w ho increasingly seek flexible w ork sche dules) as well as to org an iza tio ns '
greater needs for flexibility (see Pfeffer and Baron, 1988).
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 367/375
35
uture Trends in Labor Market Structures
Some of the key trends that are likely to have a major impact on future
labor m ark et s tructure s in m an y industr ia l natio ns are alread y clear . The se
changes in labor markets ref lec t both manager ia l /organiza t ional impera
t ives and the cha ng ing na tu re of the labor force, an d are occurr in g at both
organizational and societal levels . To a considerable extent, labor market
s t ruc tures wi l l va ry am ong count r ies , dep end ing on a na t ion ' s em ploy m ent
and training policies , employer s trategies , unionization and structure of
industr ial relat ions, economic condit ions, posit ion in the world economy,
welfare s tate s tructu res , an d so on.
Continued N eed for Functional Flexibility
Fun ctional f lexibili ty w ill con tinue to be imp orta nt to the m od er n co rpo
ration , s ince f irms mu st hav e a trained a nd a dap table w orkforce to be able
to adjust to rapid chan ges in technolog y an d bu sine ss con dit io ns. In the
United States and in most industr ial nations, then, there is l ikely to be a
continued emphasis on practices that facil i tate a multi-skil led workforce
(such as em ployee involvem ent , to ta l qual i ty m ana gem ent , job rota t ion,
an d re t ra ining) . T here should thus be cons iderable pressures on organiza
t ions to ad op t a salar ied m od el of internal labor m arke ts in bo th blue-collar
and w hi te-col lar occupat ions , wi th a cor resp ond ing decl ine in the ind us
tr ial m od el of internal labor m ark ets (see O sterm an , 1988).
Techno logical advan ces and th e con tinue d nee d for functional f lexibil ity
are l ikely to increase the impo rtan ce of edu catio n, a view that is consis ten t
w ith m ain stre am theories of post- in dustr ia l ism (e.g., Bell, 1976). Ed uca tion
wil l becom e perhap s the mo st impor tan t de term inan t of job placem ent and
life chances, and the gap separating high and low educated people will
gro w . This will increasingly div ide labor force inside rs f rom ou tsid ers , an d
ins iders am on g themselves . The emph as is on t ra ining (and re-t ra ining) a lso
underscores the necessity to s tudy other sources of skil ls in addit ion to
formal educat ion, such as company t ra ining, vocat ional apprent iceship
systems, and so on.
Continued Need for Num erical Em ployment Flexibility
Organiza t ions ' needs for numer ica l /employment f lexibi l i ty are a lso
likely to con tinu e into the future. Recessionary busine ss con dit ion s and the
chal lenges in t rod uce d by global comp et i t ion hav e led man y emp loye rs to
do w nsiz e the ir workforces in orde r to s tay in bus iness a n d /o r to m ainta in
profits. Bu nn ing (1990), for ex am ple , rep ort s that 35 % of 1200 pe rso nn el
m an ag ers in the Un ited States in a 1988 surv ey w ork ed for comp anie s th at
ha d dow nsiz ed du r ing the pas t twelve m on ths . This do w nsiz in g largely
repre sen ts del aye r ing , or a f lattening of the corp orate py ram id du e to a
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 368/375
351
reduction in midd le-managem ent positions (Cheek and C am eron, 1990).
As a result of dow nsizing and patterns of job creation, employm ent in the
United States is con tinuing to shift toward middle - and small-sized firms
(see Granove tter, 1984). These trends are also observed in E urope , though
there employers generally have less wage and emp loyment flexibility due
to constraints imposed by collective barga ining agreem ents and other laws
governing industrial and labor relations (but see Note 4).
Growth of Inflexible Labor
At the same time that employers are increasingly likely to need a m ore
flexible workforce, labor force trends in the United States are tending to
produce a more inflexible labor force. For example, the average age of the
Am erican labor force is likely to increase, as the baby boom cohort ages and
a smaller number of young persons from the baby bust cohort enters the
workforce (Johnston and Packer, 1987). The labor force thus will be made
up m ore of older m en and wom en w ho are less mobile because they have
considerable experience with firms and occupations, and who are less
willing and able to be retrained for new
jobs.
In addition, wom en w orkers
may mo re often seek perm anen t employm ent in the future, m aking them
less available for pa rt-time and temporary work (see Osterm an, 1988).
These labor force trends will constrain employers' ability to achieve
functional as well as num erical/em plo ym ent flexibility. These trend s also
poin t to the need for research on the changing labor force. For
example,
the
combination of an aging workforce and decline in long-term employment
raises important questions rega rding pattern s of retirement and portability
of pensions. (See the discussion on life course changes by O 'Rand, this volume.)
Future Labor Market Structures: Two Scenarios
While these three trends are likely to help shape the labor market
structure s of the future, their exact contou rs cannot be predicted with a high
deg ree of confidence. There are both optimistic (upg rading) and pessimis
tic (polarization) scenarios that might describe these future labor market
structures.
Upgrading of fobs and Skills Organizations' needs for functional
flexibility—coupled with an increased emphasis on education— may result
in massive upg rad ing of the skill levels of organizations and occupations,
This is reminiscent of Bell's
1976)
description of
the
post-industrial know l
edge society. Piore and Sabel 1984:279) describe this scenario as resembling
the artisans' republic envisioned by Rousseau and Prou dho n. In this
scenario, employees will have fairly high levels of job security, and there
will be a decline in secondary labor marke t jobs.
One way by which this scenario might be implemented is through the
developm ent of netwo rk organizations. The network organization is
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 369/375
352
character ized by rec iprocal pa t tern s of com mu nicat ion and ex chang e and is
an a l ternat ive gove rnance s t ruc ture to m arkets an d hierarchies ( see Powel l ,
1990). Ne tw ork organiza t ions represent c lus ters of do w nsize d, focused
bus in ess uni ts coordinated by market mec han isms ra ther tha n by layers of
m idd le-m ana ge m ent p lann ers and schedulers (Snow et a l. , 1991) . These
networks take var ious forms, inc luding consor t iums (e .g . , the Japanese
Keiretsu
joint ventures , and contractual par tner ing of different kinds
des ign ed to extend an org aniz ation 's capabil i t ies . These pa tte rns are
a l ready co m mo n in m any in dus tr ies in the Un i ted S tates, inc lud ing: con
s t ruct ion, boo k publ ishing , film a nd recording , ban king , auto s , te lecom mu
nica t ions and c om pute rs . They are a lso represented by the Emil ian m od el
in Italy and the Japanese textile industry (Powell, 1990).
The netw ork o rganiza t ion is one possible solut ion to the orga niza t io n 's
ne ed s for both functional and num er ic a l /e m plo ym en t flexibil ity. N etw ork
orga niza t io ns ten d to be re lat ively smal l and thu s have the capaci ty to
achieve higher levels of functional flexibility than is feasible in larger
organ iza t ions . F irms tha t par t ic ipa te in such netwo rks wi l l hav e a smal l
permanent core , and increas ingly use subcontrac tors such as consul t ing
f irms to pro vid e no n-p erm an ent em ployees . IBM, for exam ple , has been
ve nd ori ng lower- level services (such as intern al cafeter ias an d m ail roo m s)
for som e years , an d has also sp un off higher- level functions, such as m an y
activit ies associated w ith em plo ym en t, that are carr ied out by an aff il iated
company. These non-permanent employees ( f rom the point of v iew of
IBM's commitment to them) might themselves be re la t ive ly permanent
m em be rs of their con sultan t f irms. Ov er fifteen p ercent of w or k o rga niz a
t ion s in the U nited States in 1991 sub-con tracted o ne or m ore occ up ation al
activities (see Kalleberg and Schmidt, 1996).
Ac hievin g this optim istic scenario s likely to requ ire som e state interv en
t ion. A nat iona l em ploy m ent pol icy appea rs to be necessary in order to
pro vid e ade qua te t ra ining, which is a precon di t ion for up gra din g in ma ny
kin ds of occupa tions. Public policies are also ne ed ed to facil itate em plo y
m en t security, wh ich is a key co m po ne nt of the salar ied m od el of intern al
labor marke t s . European count r ies such as Ge rma ny and S wed en have ha d
qu ite effective em plo ym en t policies in this rega rd; by contrast , the U nite d
Sta tes t radi t ional ly ha s lacked such an emplo ym ent pol icy (see O sterm an,
1988).
These points ha ve no tb ee n lost on the Cl inton Ad min is t ra t ion, w hich
has em ph as iz ed the impor tanc e of re t ra ining and of em plo ym en t secu r i ty ,
rather than job security.
The emergence of ne twork organiza t ions sugges ts the need to s tudy
careers as they unfold betw een organ iza t ions , not jus t wi thin th em . Job
search and informat ion networks become more sa l ient i f workers and
m ana ge rs need to change organ iza t ions mo re often. F lows of hu m an capi tal
am on g organiza t ions a lso need to be inves t iga ted. For exam ple , we kno w
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 370/375
5
very l i t tle abo ut source s of t raining other than form al educ ation or on-the-job
tra in ing. M oreov er , pa t te rns of sub-contrac t ing a nd o ther forms of cont in
gen t em ploy m ent am ong ne tw orks of o rgan iza t ions rema in l a rge ly un ex
plored.
Polarization between Core and Periphery On the other hand, organiza
t ions ' e ffor ts to achieve grea ter functional an d nu m er ica l /e m pl oy m en t
flexibility m ig ht re sult in a po lariz atio n of the labo r force into a sm all gr o u p
of ful l- t ime, functionally f lexible permanent core workers (who have au
tonomy, a great deal of human capital , and often high trust relat ions with
the i r employers ) and a la rge group of per iphera l workers (who may be
re la tive ly unski l led a n d /o r outs ide rs to the labor
marke t ) .
These pe r iph e ra l
workers represent a modern, pos t - indus t r ia l vers ion of Marx ' s reserve
a rm y of the une m ploy ed : they a re s t and by o r d i sposab le wo rke rs w ho
are avai lable when needed and discarded when they are not (see a lso
B ra v e rm a n ,
1974;
M ichon , 1981). H arriso n (1994) refers to these pe rip he ral
w or ke rs as the d ar k s ide of f lexible prod ucti on . Piore and Sabel (1984:279)
descr ibe th is regim e as resemb l ing .. . the o ld Bourbo n kin gd om of Nap les ,
where an is land of c raf tsmen, producing luxury goods for the cour t , was
surrounded by a subprole tar ian sea of misery .
There are at least two versions of this pessimistic scenario (see, e.g. ,
Esp ing -A nd erse n, 1993). W ithin a natio n, there can be a polariz at ion
betw een a sm al lco re of h ighly up gra de d core ins iders , and e i ther: (1) a la rge
service proletariat consisting of lots of bad jobs; or (2) a relatively few bad
jobs an d a large pop ula t ion of labor force outs id ers . T he typ e of po lariza t ion
that occurs de p en d s on factors specific to cou ntry , inclu ding : the role of the
welfare s tate (such as the opportunit ies provided by the welfare s tate for
ext ra- labor m ark et sup por t ) ; barr ie rs to enter ing the labor m arket ; and the
system of ind ust rial relat ions. In the Un ited States , for exa m ple, the g ro w th
of a low wage labor force in the weakly unionized service sector would be
l ike ly , if i t w ere no t for the cons t ra in ts on such dev elop m ent im po sed by the
lack of persons who are wil l ing to accept relat ively low paying, unstable,
an d oth erw ise inferior jobs (O sterm an, 1987). (On the othe r ha n d, the s tu dy
by Alexa nde r an d Entw is le , th is volu m e, sugge s ts tha t there m ay be m ore
such low ski l led and educated persons than we might have assumed. )
This scenario suggests greater urgency for certain kinds of s tudies of
careers . For exa m ple , research is ne ed ed on the corre la tes of u ne m plo y
ment. As in the early research on dual labor markets , i t would also be
imp or tan t to s tud y the cond i t ions und er ly ing and pe rp e tua t ing secondary
labor market work and the barr ie rs tha t h inder secondary workers f rom
enter ing the pr i m ary labor m arke ts . Again , s tudie s of the sources of
education and training, and the transi t ions from the s i tes where skil ls are
tau gh t to the workp laces w he re they are u t i l ized, a re l ike ly to take on ad de d
impor tance .
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 371/375
54
It is too early to tell whether the optimistic or pessimistic scenario will
more accurately describe future labor market structures in the various
industrial nations. Nevertheless, there are several reasons to suspect that
the former, more optimistic scenario might be more appropriate in many
industrial coun tries. A num ber of factors constrain the expansion of a low
wage labor force in industrial societies; these range from the considerable
power of unions in many European countries, to national employment
policies in countries such as Germany and Sweden, to the unavailability of
persons willing to work in secondary m arket
jobs.
Increased international
com petition may also encourage firms
to
remain small and highly function
ally flexible. Evidence from the
U.S.,
U.K., Canada, Germany , No rwa y and
Sweden indicates that ...everyw here, the trend favors the higher-grade
occupations such that the shape of the post-industrial occupational hierar
chy is biased toward the top and the middle, rather than the bottom
(Esping-Andersen et al., 1993:53).
These two scenarios are of course oversimplified, an d these ideal types
need further development and refinement. One complication
is
introduced
by the trend toward internationalization of production. What is good for
one country may not be so good for
another.
he outsourcing of semi-skilled
jobs from the U.S. to other coun tries, for exam ple, is optimistic from the
point of view of the country gaining jobs, but pessimistic from the stand
poin t of comm unities in the United States that may lose jobs.
In any event what
is
clear
is
that changes
in
labor market structures pose
new empirical questions that constitute an agenda for future research.
Studies of these issues require com parative, cross-national d ata on wo rk
organ izations and their employees that are based on representative sam ples
dra wn from diverse populations. These studies need to be com plemented
by investigations of specific industries, organizations and occupations.
Research also needs to focus on netw orks of organizations, to explain how
em ploye rs interact with other organizations to obtain and train new work
ers.
Such studies are necessary to understand better the linkages between
work organizations and broader labor market contexts.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 372/375
55
Notes
1. Segm entation describes a general approach to studyin g labor markets that
goes beyond the du al framework. Labor market segmentation theory posits the
existence of distinct labor m arkets that differ in their processes of earnin gs d eterm i
nation an d other outcomes. In this view, dual labor market theory is a specific
example of labor market segm entation theories.
2.
By contrast, the traditional industrial internal labor market model organizes
work into a series of narrowly defined jobs that are tied to clear duties and
responsibilities (thereby obtaining high num erical/em ploym ent flexibility bu t low
functional flexibility) (see Osterman, 1987; 1988).
3. Such externalization of employment represents a retreat from internalized
bureau cratic emp loyment relations to a set of ongoing transactions with the o pen
labor market (Pfeffer and Baron, 1988). This barkens back to an earlier era in
economic organization when factory labor was not continuously employed but
rather subco ntracted from the outsid e (W illiamson, 1985).
4. The United States legal context is generally perceived as being m ore flexible
than m any Europ ean nations in its policies regarding layoffs. How ever, the greater
flexibility of U.S. employers in this regard is often exaggerated: th e U.S. courts have
been increasingly restrictive in their interpretations of the employment-at-will
doctrine, and this has m ade U.S. employ ers very cautious about exercising it In 1986
in California, for
example,
plaintiffs won 78% of wron gful-discharge cases with an
average total awa rd of $424,527 (Fulmer and Casey , 1990). A s a consequence, it has
been argued that the degree of employment and wage flexibility in the U.S. is not
qualitatively different from that found in Europe (Piore, 1986).
5.
It is the temporary workers—more than the regular workers—who provide
what little formal definition there is to the boundaries of the permanent employm ent
system in Japan, since it is the temporaries—not the regulars—who are issued
emp loyme nt contracts clearly delineating the fixed-term natu re of their attachment
to the firm. The depend ence on an immense qu antity of peripheral labor is a prior
condition for w hy the Japanese are able economically to implement m any of the
organization structures that promo te comm itment (Tokunaga,
1984). Some
of these
peripheral workers were formerly regular, perm anen t employees, wh o we re
hired back as tempora ries at lower wages after their retirement. ) However, the
negative features of the du al structure of Japanese indu stry are often overstated:
subcontractors in Japan are relatively independent from the larger, parent firms—
at least in indu stries such as machine tools—and em ployees of subco ntracting firms
are often highly skilled and relatively well paid (see Friedman, 1988; Ch almers,
1989).
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 373/375
356
R e f e r e n c e s
A lthaus er, Robert P. 1989. Internal Labor M ark ets. Annual Review of Sociology
15:143-161.
Alth ause r, Robert P. and A rne L. Kalleberg. 1981. Firm s, O ccu pat io ns an d the
Stru cture of Labor Ma rkets: A Co ncep tual An alysis Pp . 119-149 in Ivar Berg
(ed.) , Sociological Perspectives on abor M arkets (New Y ork: Acad emic P ress) .
Atch ison, Th om as J. 1991. The Em ploy m ent Relat ionship: Un -t ied or Re-t ied.
Academ y of Mana gement Executive 5:52-62.
Atk inson, John. 1987. Flexibi li ty or Frag m entat ion ? Th e Un ited K ing do m La bou r
Market in the Eight ies. abour and Society 12:88-105.
A ver itt , Robe rt T. 1968. The Dual Economy: The Dynamics of American Industry
Structure. N ew York : No rton .
Baron, Jam es hi. Alison Davis-Blake, and W il liam
T
Bielby. 1986. Th e Str uc tur e of
Oppor tun i ty : How Promot ion Ladde rs Vary Wi th in and Among Organ iza
t ions . Adm inistrative Science Quarterly 31:248-273.
Bell, D anie l. 1976.
The Coming of Post-Industrial Society.
N ew York: Basic Books.
Berger, Suzanne and Michael J. Piore. 1980. Dualism and Discontinuity in Industrial
Societies. Cambridge : Cambridge Univers i ty Press .
Blau, Peter M . and O tis Dudley D unca n. 1967. The Am erican Occu pational Structure.
New York: Wiley.
Bluestone, Barry. 1970. The Tripart i te Econom y: Labor M arkets and the Wo rkin g
Poor . Poverty and Hum an Resources Abstracts 5:15-35.
Boyer, Rob ert. 1987. La bou r Flexibili t ies: M any Form s, U nce rtain Effects. Labour
and Society
12:107-129.
Boyer, Robert (ed.) 1988. The Search for Labour Market Flexibility: The European
Econom ies in Transition.
Oxford : Clarendon Press .
Braverm an, Ha rry . 1974. Labor and Monopoly Capitalism. N ew York : M onth ly
Review Press.
Bu nning , Richard L. 1990. The Dy nam ics of D ow nsiz ing. Personnel Journal 69:
68 75.
Chalmers , Norma J . 1989 .
Industrial Relations in japan: The Peripheral Workforce.
London: Rout ledge .
Cheek , Ge ra ld D. and Wal ter Cam eron . 1990. Cor pora te Do wn siz ing : An Am eri
can Trainer's Affliction. Industrial and Com mercial Training 22:26-31.
Co hen , Yinon and Jeffrey Pfeffer. 1986. O rganiza t ional Hiring Stan da rds .
Admin
istrative Science Quarterly 31:1-24.
E dw ar ds , Richard. 1979. Contested Terrain. N ew York: Basic Books.
Esp ing-A nder sen, G0sta. 1993. Post-Indu strial Class Structure s: A n An alyt ic
Fra me wo rk . Pp . 7 -31 in Ges ta Esp ing-A ndersen (ed .) , Changing C lasses: Strati
fication and Mobility in Post-Industrial Societies. London: Sage.
Esping-A ndersen , Gasta , Zina Ass imak opoulou and Kees van Kersbergen . 1993 .
Tren ds in Con tem porary C lass St ruc tura t ion : A Six-na tion Co m paris on . Pp .
32-57 in Ge sta Es ping-A nders en
(ed.) ,
Changing Classes: Stratification and Mo bility
in Post-Industrial Societies.
London: Sage.
Freedman, Marcia K. 1976. abor Markets: Segments and Shelters. Mo ntclair , N ew
Jersey : Al lanheld , O sm un.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 374/375
357
Friedman, David. 1988. The
Misunderstood
Miracle: Industrial
Development
and
Political Change in
Japan. Ithaca, New York: Cornell University Press.
Fulmer, William E. and Ann
Wallace
Casey. 1990. Employment
at Will:
Op tions for
Managers.
The Academy o f Management Executive.
4:102-107.
Gordon , David M. 1972.
Theories
of
Poverty
and
Underemployment.
Lexington,
Massachusetts.: D C Heath.
Granovetter, Mark. 1984. Small is Bountiful: Labor Markets and Establishment
Size. American Sociological Review 49:323-334.
Handy, Charles. 1989. The Age of Unreason. Boston: Harvard Business School Press.
Harrison, Bennett. 1994.
Lean and Mean: The Changing Landscape of Corporate Power
in
the Age
o f
Flexibility.
New York: Basic Books.
Johnston, William B. and A rnold H. Packer.
1987. Workforce 2000: Work and Workers
for the 2 st Century.
Indianapolis: Hudso n Institute.
Kalleberg, Ante L. and Kathryn Schmidt. 1996. Contingency Workers and
Contingent W ork: Part-Time, Temporary, and Subcontracting Employm ent
Relations in U.S. Organ izations. In Arne L. Kalleberg, David Knoke, Peter V.
M arsden, and Joe L. Spaeth,
Organizations in Am erica: A Portrait of their Structures
and Human Resource Practices.
New bury Park: Sage.
Kalleberg, Arne L. and Aage B. Serensen. 1979. The Sociology of Labor M arkets.
Annual Review of Sociology
5:351-379.
Koike, Kazuo. 1978. Japan's Industrial Relations: Characteristics and Problem s.
Japanese Economic Studies
7:42-90.
Lane, Christel. 1989.
Management and Labour in Europe: The Industrial Enterprise in
Germany
Britain and
Prance. Brookfield, VT: Gow er.
M angum , Garth, Donald Mayall, and Kristin N elson. 1985. The Temporary Help
Industry: A R esponse to the Dual Internal Labor M arket.
Industrial and Labor
Relations Review 38:599-611.
M eulders, Daniele and Luc Wilkin. 1987. Labour Market Flexibility: Critical
Introduction to the Analysis of a C oncept.
Labour and Society
12:3-17.
Michon, F rancois. 1981. Dualism and the French Labour
Market:
usiness Strategy,
Non-standard Job Forms and Secondary Jobs. Pp . 81-97 in Frank Wilkinson
(ed.)
The Dynamics o f Labour Market Segmentation.
New York: Academic Press.
Osterm an, Paul.
1987.
Turnover, Employment Security, and the Performance of the
Firm . Pp . 275-317 in
M.M.
K leiner, R.N. Block, M. Roomkin, S Salsburg (eds.),
Human Resources and the Performance
of
the
Firm. Madison, Wisconsin: Industrial
Relations R esearch Association.
Osterman, Paul. 1988.
Employment Futures: Reorgan ization Dislocation and Public
Policy.
New York: Oxford University P ress.
Pfeffer, Jeffrey and Jam es N. Baron. 1988. Taking the Workers Back Out: Recent
Trends in the Structuring of Employm ent. Pp . 257-303 in Barry M. Staw and
Larry L. Cummings (eds.),
Research
in
Organizational Behavior
Vol. 10. Green
wich,
CT:
JAI Press.
Pfeffer, Jeffrey and Yinon Cohen. 1984. Determina nts of Internal Labor Markets in
Organizations.
Administrative Science Quarterly
29:550-572.
Piore, Michael J. 1986. Perspectives on Labor Market Flexibility.
Industrial
Relations
25:146-166.
8/16/2019 Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)
http://slidepdf.com/reader/full/alan-c-kerckhoff-generating-social-stratification-toward-a-new-research-agenda 375/375
358
Piore, M ichael
J
and Charles
F.
Sabel. 1984. The Second Industrial Divide: Possibilities
for
Prosperity
New York: Basic Books.
Pollert, Anna. 1988. The 'Flexible
Firm':
Fixation or
Fact?
Work
Employment and
Society
2:281-316.
Powell, Walter W. 1990. Neither Market Nor Hierarchy: Ne twork Forms of
Organization.
Research
in
Organizational Behavior
12:295-336.
Reich, Michael. 1984. Segm ented Labour:
Time Series
Hypothesis and Evidence.
Cambridge Journal o f Economics
8:63-81.
Rosenberg, Sam. 1989. From Segm entation to Flexibility.
Labour
and Society
14:363-407.
Sengenberger, Werner. 1981. Labour Market Segmentation and the Business
Cycle. Pp . 243-259 in Frank Wilkinson (ed.),
The Dynamics of Labour Market
Segmentation
London: Academic Press.
Sengenberger, Werner and
Ch.
Kohler.
1987.
Policies of Workforce Reduc tion and
Labour M arket Structures in the Am erican and Germ an Automobile Industry.
Pp. 245-269 in Roger Tarling (ed.),
Flexibility in Labour Markets
Orlando: Aca
dem ic Press.
Snow, Charles C , Raymond E. Miles and Henry J Coleman, Jr 1991. M anaging 21st
Century Organizations.
Organizatiotwl Dynamics
20,4:5-20.
Spilerman , Seymour. 1977. Careers, Labor Market Structure, and Socioeconomic
Achievement.
merican Journal
of
Sociology
83:551-593.
Tokunaga, Shigeyoshi. 1984. The Structure of the Japanese Labor M arket. Pp.
25 55 in Shigeyoshi Tokunaga and Joachim B ergmann (eds.),
Industrial Relations
in Transition: The Cases of Japan and the Federal Republic of Germany
Tokyo:
University of Tokyo Press.
Williamson, Oliver E. 1985.
The Economic Institutions of Capitalism
New York: Free
Press.