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



Generating Social Stratification

Toward a N ew Research Ag end a



This page intentionally left blank



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



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



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



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



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



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



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



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



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.



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



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.



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.



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.



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,



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



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 .



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,



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.



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



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



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



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-



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



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



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.


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.



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.



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.



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.



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.



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



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 entre 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



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



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



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 entre 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



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



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



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



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

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-



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 entre 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



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-



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.



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.



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


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 .



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.



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.



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.



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

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



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



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



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



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



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



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



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



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



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

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



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



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.



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



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.



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.



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.



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.



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.



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 ,



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.



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



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).



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



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



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.



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



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,



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 .



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

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-



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 .



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



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


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.



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.



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 .


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 .


48:838-851.

H ey ns , Barb ara. 1974. Social Selection an d Stratification w ithi n Schools.

American


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.


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.



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


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.


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


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.


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.


54:181-195.



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



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



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

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



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

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



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

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



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



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



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

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

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



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



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



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



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.



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)



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



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***



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**



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-



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



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



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 ,



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 .



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.


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.



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.


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.



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


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.



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?


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.



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.



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



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. ,



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,



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,



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



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



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



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



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



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



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

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 .



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.



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.



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-



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



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.



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.



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.



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.



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.



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.


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



229

ndependent Variables

Femaie


Professional I

Professional II


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 .



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.



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


Professional I

Professional II


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



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



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-



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



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 .



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 .



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



P

e

c

n

a

g

o

P

P

a

o

r

o

O

C

O

N

>

O

0

1

—

I

—

I



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



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,



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



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



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


Annual Increase in Segregation


SES Segregation Index


An nua l Increase in Segregation


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



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,



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



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



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



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.



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.



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.


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.



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.



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



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.



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



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-



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.



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.



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



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



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



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,



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



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



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



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,



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.



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.


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.


67(2):84-89. And One

more thought. p.91.



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.



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-



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.



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.



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



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



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



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 ,



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.



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



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

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.



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



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-



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-



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



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



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-



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



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)


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)


alter nativ e 19.0 25.3 55.7 100.0

(114)

Gamma=.686 p<.001



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 .



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



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



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



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.



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)



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



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.



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


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.


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


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.



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.



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.



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



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



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



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



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 -



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



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-



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



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



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



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-



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



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


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-



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



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


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



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


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



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,



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.



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.



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.


(forthcoming).

Thurow , Lester

C

1975. Generating

Inequality:

Mechanisms of Distribution in the U.S.

Economy. NY: Basic Books



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-



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



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.



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



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.



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.



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-



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



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,



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



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



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



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.



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



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



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



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



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?



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



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.



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



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



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



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



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.



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



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 -



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



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.



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.



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.


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.



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 .



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.



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



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;



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



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



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.



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-



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



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-



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-



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 ,



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



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



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



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



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



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-



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.



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



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



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



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



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 -



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.



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



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

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.



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



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-



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



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



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.



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.



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.



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-



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.



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



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-



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



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.



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



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.



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



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



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



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



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



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).



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



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



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.



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



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



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.



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



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.


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.



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.



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


. 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.



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.



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 .


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.


52:482-494.



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 .



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



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



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



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



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—



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



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



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



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


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)



College Major: '


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.



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.



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



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



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-



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



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



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



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



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



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



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



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



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.



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



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.



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.


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.


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.



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.



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



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



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 ,



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-



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



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).



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



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



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



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 .



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.



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).



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.



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.


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.



358

Piore, M ichael

J

and Charles

F.

Sabel. 1984. The Second Industrial Divide: Possibilities

for

Prosperity


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.

Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)

Documents

Transcript of Alan C Kerckhoff-Generating Social Stratification_ Toward a New Research Agenda (Social Inequality)