ED 295 823

TITLE

INSTITUTION

REPORT NOPUB DATENOTE

AVAILABLE FROM

PUB TYPE

EDRS PRICEDESCRIPTORS

DOCUMENT RESUME

SE 049 204

Demographic Trends and the Scientific and EngineeringWork Force. A Technical Memorandum.Congress of the U.S., Washington, D.C. Office ofTechnology Assessment.OTA-TM-SET-35Dec 85153p.; Graphs, charts and printing on colored pagesmay not reproduce well.Superintendent of Documents, U.S. Government Printiug,iffice, Washington, DC 20402 ($6.00).Reports - Descriptive (141)

MF01/PC07 Plus Postage.*College Science; *Demography; *EngineeringEducation; *Engineers; Equal Opportunities (Jobs);Higher Education; *Labor Force; Labor Market;*Professional Education; Professional Personnel;Researchers; Science Careers; Science Education;Scientific Personnel; Scientists

ABSTRACTThe Federal Government has acknowledged its key role

in educating and assuring an adequate supply of scientists andengineers since World War II. Although scientists and engineersrepresent only three percent of the national work force, they areconsidered by many to be a crucial element in the nation's efforts toimprove its economic competitiveness and national security. As aresult of this study, it was found that a decline in the college-agepopulation in the next decade and an increase in the fraction of the18 to 24-year-old cohort that will be drawn from minority populationscould affect the supply of scientists and engineers. In general,there does not seem to be evidence for concern that demographictrends will lead to shortages. This memorandum contains thefollowing: (1) "Executive Summary"; (2) "Overview and Findings"; (3)"Demographic Trends: Undergraduate and Graduate Education"; (4)"Demographic Trends and the Academic Market for Science andEngineering Ph.D.s"; (5) "The Industrial Market for Engineers"; and(6) "Demographics and Equality of Opportunity." Also two appendicesare included. One discusses the education and utilization ofbiomedical research personnel. The second lists responses to a surveyon factors affecting minority participation in science andengineering. (RT)

***********************************r***********************************

Reproductions supplied by EDRS are the best that can be madefrom the original document.

***********************************************************************

,

,..

Office of Technology Assessment

Congressional Board of the 99th Congress

TED STEVENS, Alaska, Chairman

MORRIS K UDALL, Arizona, Vice CI-airman

Senate

ORRIN G HATCHUtah

CHARLES McC MATHIAS, IRMarylana

EDWARD M KENNEDYMassachusetts

ERNEST F. HOLLINGSSouth Carolina

CLAIBORNE PELLRhode Island

WILLIAM J PERRY, ChairmanH&Q 7-, i7nology Partners

DAVID S POTTER, Vice ChairmanGeneral Motors Corp (Ret )

EARL BEISTLINEConsultant

CHARLES A 3OWSHERGeneral Accounting Office

House

GEORGL E. BROWN, JR.California

JOHN D. DINGELLMichigan

CLARENCE E. MILLEROhio

COOPER EVANSIowa

JOHN H. GIBBONS(Nonvoting)

Advisory Council

CLAIRE T. DEDRICKCalifornia Land Commission

JAMES C. FLETCHERUniversity of Pittsburgh

S DAVID FREEMANConsultant

GILBERT GUDELibrary of Congress

Director

JOHN H GIBBONS

DON SUNDQUISTTennessee

MICHEL T. HALBOUTYMichel T. Halbouty Energy Co.

CARL N. HODGESUniversity of Arizona

RACHEL McCULLOCHUniversity of Wisconsin

LEWIS THOMASMemorial Sloan-Kettering

Cancer Center

This is an OTA Technical Memorandum that has been neither reviewed norapproved by the Technology Assessment Board.

3

1 DEMOGRAPHIC TRENDSAND THE SCIENTIFICAND ENGINEERING

WORK FORCE

A TECHNICAL MEMORANDUM

DECEMBER 1985

Technical Memoranda are issued by OTA on specific subjects analyzed in re-cent OTA reports or on projects in process at OTA. They are issued at the re-quest of Members of Congress who are engaged in committee legislative actionswhich are expected to be resolved before OTA completes its assessment.

4

CONGRESS OF THE UNITED STATES

CM* el Technology AssessmentWashing len. 0 C 20:00

Recommended Citation:U.S. Congress, Office of Techno' ogy Assessment, Demographic Trends and the Scientificand Engineering Work ForceA Technical Memorandum (Washington, DC: U.S. Govern-ment Printing Office, OTA-TM-SET-35, December 19P5).

Library of Congress Catalog Card Number 85-600622

For sale by the Superintendent of DocumentsU.S. Government Printing -iffice, Washington, DC 20402

Foreword

Scientists and engineers represent only 3 percent of the national work force, butthey are crucial to the Nation's efforts to improve its economic competitiveness andnational security. Their work makes possible advances in knowledge that improve rkewell-being of people all over the world.

This technical memorandum explores the effect that char.ges in the size and com-position of the U.S. population may have on the science and engineering work forceThe study was requested by the Science Policy Task Force of the House Committeeon Science and Technology.

OTA concluded that while demographic trends are important and should be con-sidered in thinking about future scientific output, they are not a controlling variable.The flexibility of the American education system and the adaptability of our scientistsand engineers enable market forces and career choices to play a dominant role in balancingsupply with demand. OTA also concluded that it is important to maintain and strengthenexisting efforts to improve access to scientific and engineering careers for disadvantagedgroups.

1(141444.JOHN H. GIBBONSDirector

'I!

Advisory Group

Jemer A. HaddadConsultant

Lee HansenUniversity of Wisconsin

Lilli S. HornigWellesley College

Luther S. WilliamsAtlanta University

Dael Wolf leUniversity of Washington

Workshop on the Education and Employment ofScientists and Engineers at the Nation's Colleges and Universities

Sue BerrymanThe Rand Corp.

Radford Byer lyCommittee on Science and TechnologyU.S. House of Representatives

Lloyd Cooke, Consultant

Michael CrowleyNational Science Foundation

Charles DickensNational Science Foundation

Daniel DruckerUniversity of Florida

Alan FechterNational Research Council

Jerrier A. HaddadConsultant

Maria HardyLouisiana State University

Allan HoffmanNational Academy of Sciences

John HolmfeldCommittee on Science and TechnologyU.S. House of Representatives

Lilli S. HornigWellesley College

Fred LandisUniversity of Wisconsin

Iv

W. Edward LearAmerican Society for Engineering Education

Henry LowendorfYale University

Shirley MalcomAmerican Association for the Advancement of

Science

Shirley Mc BayMassachusetts Institute of Technology

Michael McPhersonBrookings Institution

Denis F. PaulState Education DepartmentState of New York

Willie PearsonWake Forest Uni versity

Geraldine RichmondUniversity of Oregon

Harrison ShullUniversity of Colo; ado

F. Karl Wilt ,rockSouthern Methodist University

Luther S. WilliamsAtlanta University

Dael WolfeUniversity of Washington

7

OTA Project Staff for Demographic Trends andthe Scientific and Engineering Work Force

John Andelin, Assistant Director, OTAScience, Information, and Natural Resources Division

Nancy Carson Naismith, Science, Education, and Transportation Program Manager

Eugene Frankel, Project Director

Kathi Hanna Mesirow, Analyst

Barbara Berman, Research Assistant

Jonathan Atkin, Research Assistant

Marsha Fenn, Administrative Assistant

Betty Jo Tatum, Secretary

Chris Clary, Secretary

Gala Adams, Clerical Assistant

Contractors

Michael Mandel National Academy of Sciences VV. Curtiss Priest

Scientific Manpower Commission Robert G. Snyder

Other OTA Contributors

John Aiic Eric Basques Philip P. Chandler, II Lirda Garcia

Barry Holt Marcel C. La Follette

Special Thanks for Extraordinary Assistance To:

Alan Fechter, National Research Council

Shirley Malcom, American Association for the Advancement of Science

Betty Vett, Scientific Manpower Commission

0

Reviewers and Other Contributors

Josefina Arce de SanabiaUniversity of Puerto Rica

Alexander AstinUniversity of California at Los Angeles

Elinor BarberInstitute of Inteinational Education

Pedro BarbosaUniversity of Maryland

Dennis ChamotAFL-CIO

Susan ChipmanOffice of Naval Research

Juanita J. CollinsCrown-Zellerbach

Ewaugh FieldsUniversity of the District of Columbia

Michael FinnOak Ridge Associated Universities

Robert A. FinnellNational Action Council for Minorities

Engineering

Penny FosterNational Science Foundation

Debra GeraldU.S. Department of Education

John E. GibsonUniversity of Virginia

Edward GlassmanU.S. Department of Education

Howard GobsteinU.S. General Accounting Office

Manuel GomezUniversity of Puerto Rico

Richard GriegoUniversity of New Mexico

Harold HodgkinsonInstitute for Educational Leadership

vi

Dana JurakConsultant

Clara Sue KidwellUniversity of Califoraia at Berkeley

Charlotte KuhAT&T Communications

Jules LapidusCouncil of Graduate Schools

Cora MarrettUniversity of Wisconsin

Doris MerrittNational Institutes of Health

William MichaelNational Academy of Engineering

Robert NeumanAmerican Chemical Society

Sheila Pfaff linAmerican Telephone & Telegraph Co.

Don I. Phillipsin National Academy of Sciences

Vivian Pi-,nHoward University

Beverly PorterAmerican Institute of Physics

Edith RuinaMassachusetts Institute of Technology

Anne Scanley,National Academy of Sciences

Allan SingerInstitute of Medicine

Anne B. SwansonNational Science Foundation

Peter SyversonNational Research Council

Gail ThomasJohns Hopkins University

9

Contents

EXECUTIVE SUMMARY .. ....... .. .

CHAPTER 1: Ovcrview and Findings.......... . ... ... . ...... . 11

CHAT TER 2: Demographic Trends: Undergraduate and Graduate Education ... 33

CHAPTER 3: Demographic Trends and the Academic Market forScience and Engineering Ph.Ds 55

CHAPTER 4: The Industrial Market for Engineers 91

CHAP? ER 5: Demographics and Equaliiy of Opportunity 113

APPENDIX A: The Education and Utilization of BiomedicalResearch Personnel 133

APPENDIX B: Responses to an OTA Survey on Information Needs Related toMinorities in Science and Engineering 141

Page3

Executive Summary

Executive Summary

The key role of the Federal Government in edu-cating and assuring an adequate supply of scien-tists and engineers has been acknowledged sincethe close of World War II. Scientists (includingsocial scientists; see figure 1) and engineers rep-resent only 3 percent of the national work force,but are considered by many to be a crucial ele-ment in the Nation's efforts to improve its eco-nomic competitiveness and national security. TheOffice of Technology Assessment was asked bythe Task Force on Science Policy of the HouseCommittee on Science and Technology to exam-ine the implications of long-term demographictrends for scientific and engineering personnel pol-

icy, and to consider as well the barriers to andfuture trends in the participation of women andminorities in scientific and engineering careers.

OTA found two significant demographic trendsthat could affect the supply of scientists and engi-neers. The first is a decline iii the college-age pop-ulation in the next decade. The number of 18 to24 year olds will drop from a peak of 30 millionin 1982 to about 24 million in 1995, or by 22 per-cent. Labor market specialists estimate that thedecline in 18 to 24 year olds could leld ko a dropin college enrollments of 12 to 16 percent betweennow and 1995.

Figure 1 Population of U.S. Scientists and Engineers bi Field, 1983

Mathematicalscientists

3% \ilk

Aeronautic illastronautical

2%Chemical

.4............- 3%

Scientists44%

SOURCE National Science Foundation, Screnre and Technology 9: ---,,,-, NSF 85-305, January 1985, pp 53-85

4

Second is an increase in the fraction of the 18to 24 year old cohort that will be drawn from mi-nority populations, including blacks, Hispanicsand Asian-A:neric...is These groups, with the ex-ception of Asian-Americans, have historically par-ticipated less actively in science and engineeringeducation than whites.

Scienc engineering baccalaureates havecons, '.uteci 28 to 32 pea _ent of the total numberof bachelor's degrees awarded in every year since1952 (about 300,000 have been awarded annuallyin science and engineering si. -:e 1972). If thishistorical ratio continue, the expected drop in col-lege enrollments and changes in the student mixwould lead to a decline in the number of scien-tists and engineers being produced by the Nation'shigher education system. If, as has been argued,the Nation will require an increasing number ofscientists and engineers to meet its national secu-rity, economic growth, and technological inno-vation requirements, this possible decline in sci-ence and engineering baccalaureates could posea significant problem. Fortunately, OTA finds thisscenario to be unconvincing.

It is entirely possible that the supply of peopletrained in science and engineering will not decli_at all, despite the drop in the college age popula-tion. This is possible for several reasons. First, thedecline in enrollment may not be as severe as pre-dicted. The proportion of 25 to 44 year olds whoattend college, and whose population cohort willnot be decreasing, has increased dr amatically inthe past decade, especially among women. Thenumber of foreign nationals attending U.S. col-leges and universities has increased more than ten-fold ;n the past 30 years. If increases of 25 to 44year olds and foreign nationals continue over thenext decade, they could compensate, in part, farthe decline in the number of 18 to 24 year olds.

Second, there could be increases in the rate atwhich college students choose to major in scienceand engineering. Less than 7 percent of the mem-bers of a given age cohort currently receive abachelor's degree in a science or engineering field.This is near the low point of the last decade. Aslight increase in the rate of selection of scienceand engineering careers among college age stu-dents could more than compensate for the decline

in 18 to 24 year olds projected for the 1982-95 timeframe. For example, if the rate of attainment ofscience and engineering bachelor's degrees wereto increase from the 1082 level of 6.8 percent ofthe 22 year old population to the 1973 lex. ei cf8.2 percent that would cause a rise in e num-ber of degrees awarded in these fields of 20 per-cent, which would exceed the projected demo-graphic decline. Projections of shortage might wellserve to increase science and engineering enroll-ments, as the labor market responds by promis-ing increased pay or better job opportunities.

Third, science and engineering workers can beobtained by employing a higher proportion ofthose obtaining the required undergraduate de-grees, and drawing on the existing talent pool.Less than two-thirds of the science and engineer-ing baccalaureates produced in recent years haveactually become part of the science and engineer-ing work force.

Finally, on the graduate level, there appears tobe no direct relationship between the number ofPh.D.s in science and engineering, and the sizeof the graduate school age population. Between1960 and 1970 the number of full-time graduatescience and engineering sti-dents enrolled in U.S.institutions of higher education increased 150 per-cent, and the number of science and engineeringPh.D.s awarded increased 183 percent. At thesame time the population of 22 to 34 year oldsthe group which supplies the majority of gradu-ate studentsincreased only 18 percent. Between1970 and 1983, by contrast, science and engineer-ing enrollments increased by only 36 percent andannual doctorates awarded by 1 percent, but the22 to 34 year old population increased by morethan 50 percent. Accordingly, the reduction in thesize of the college-age cohort does not inevitablylead to a reduced supply of scientists and engineers.

There is nr; national market for scientists andengineers as a group. Rather, there are specificmarkets for graduates trained in particular dis-ciplines, and these markets and disciplines can ex-perience very different conditions at the sametime. For example, the National Science Founda-tion (NSF) projected in 1982 that the demand forelectrical and aeronautical engineers and for com-puter specialists could exceed the supply of grad-

; 3

uates in those fields by as much as 30 percent overthe next 5 years, while at the same time therewould be significantly fewer openings for biolo-gists, chemists, geologists, physicists, mathema-ticians, chemical, civil, and mechanical engineersthan there would be trained degree-holders. Thus,it is indi iidual disciplines, especially those linkedwith high growth or defense-oriented industries,that could experience personnel problems in thefuture; not "science and engineering" as a whole.

This finding is encouraging, because individ-ual disciplines have shown great ability to growand shrink as market conditions require. Thenumber of engineering baccalaureates, for exam-ple, more than doubled between 1976 and 1984.The number of bachelor's degrees in computer andinformational science tripled between 1977 and1982. At the same V..ne, those in education, for-eign languages, anthropology, history, and sociol-ogy all fell by 30 percent or more. Thus, collegestudents appear to be highly responsive to mar-ket signals, and appear to shift their career choicesdramatically toward fields that promise greateroccupational rewards. OTA concludes that careerchoices and market forces have a greater impacton the supply of scientists and engineers than dodemographic trends.

The possible decline in student enrollments dis-cussed above, coupled with a low retirement rateamong current faculty (most of whom were hiredin the late 1960s and will not retire until the late1990s), will lead to a very weak academic mar-ket for new Ph.D.s in the next decade. Studiescarried out by a number of analysts in the late1970s and early 1980s found that the annual de-mand for new junior faculty in the decade between1985 and 1995 would fall to less than half of thelevels experienced in the early 1970s. The annualacademic demand for new science r,,nd engineer-ing Ph.D.s could fall to as low 7.s one-third therate of hiring experienced in tin- 1970F, during the1985-95 time period. In the late 1990s the aca-demic requirements for new junior faculty in gen-eral, and for new Ph.D. scientists and engineersin particular, should increase substantially, dueto increases in both enrollments and retirements.In both c'ces the demand will most probably re-turn to the levels experienced in the early 1970s

5

and remain there through the first decade of the21st century. The pattern of decline followed byincrease will be experienced separately over thenext quarter century by each of the major scienceand engineering fields. Although these paiternsare generally acknowledged I-7 experts in the field,considerable additional research is rec ired to re-solve uncertainties in the input assumptions to theprojectionsenrollments, retirements, resigni-tions, ter, are rates, and others. The consensus ofexperts is that a combination of faculty and stu-dent demographics will lead to a weak academicmarket for new Ph.D.s over the next decade, fol-lowed by an upsurge in academic hiring between1995 and 2010.

In industry the demand for Ph.D. scientists andengineers has been strong for the past decade. Ifit should continue strong through the next de. -ade, it could compensate for the decline in the pro-jected academic demand. By the year 2000 indus-try and academia could be in stiff competition fornew science and engineering Ph.D.s.

Apart from academia, however, where knowndemographic trends exert considerable influence,projections of supply and demand for scientistsand engineers it the overall economy are fraughtwith uncertainty and are relatively unreliable.Such projections depend on assumptions aboutthe future behavior of variables such as gross na-tional product (GNP) growth, defense spending,technological change and Federal and industrialR&D ex; enditures, which themselves are not knownwith any degree of certainty. There is, moreover,no validated model that can reliably predict thecareer choices of undergraduates or graduate stu-dents, or their responsiveness to changes in de-mand in technical fields. In the absence of sucha model, analysts have in the past tended to as-sume the continuation of existing trends, an as-sumption that can lead to gross inaccuracies.Given the problems with forecasting supply anddemand for scientists and engineers, predictionsof shortages based on such forecasts should betreated with considerable skepticism.

Most projections of shortages assume tacitlythat demand and supply are independent, so thatthere can be no adjustment to supply-demand

6

gaps. This assumption makes sense, however,only under certain very restrictive conditions:

1. that demand for the final product is rela-tively unaffected by labor costs;

2. that supply is not appreciably affected bywag,- changes; and

3. that the skill shortages are unique, in thatworkers possessing them cannot be replacedby workers from other occupations or bynew technology.

If any of these conditions are violated, then aprojected supply-demand gap will be closed bymarket adjustment. On the supply side, an in-creasing number of entry-level professionals canbecome trained in the shortage area, or experi-ence,' workers from neighboring specialties canmove into the shortage occupation. On the de-mand side, employers can offer higher salaries,increase their search and recruiting efforts, re-arrange jobs to utilize available skills, educationand experience more efficiently, or make largerinvestm .its in training and retraining of their existing work force. It is more important to try tounderstand the process of adjustment of the tech-nical labor market to supply-demand imbalancesthan it is to make long-term projections.

On the supply side, we have seen that studentsshift their career interests to follow signals fromthe marketplace. However, this response is nota short-term remedy to a supply-demand gap be-cause it takes 4 years to train a new engineer and6 years (from the baccalaureate) to train a newscientist in the university. In a fast-moving mar-ket, the needs of employers may change by thetime entering students graduate.

Occupational mobility from related fields is theshort-term response to shortages. The primaryconcern about occupational mobility from the em-ployer's point of view is that it may lower thequality of work or lead to expensive retraining.From a societal point of view, however, occupa-tional mobility appears to have served the Na-tion well in meeting its needs for technical per-sonnel. The Manhattan project in World War IIthe Apllo progiam in the 1960s, the environ-mental and ert,rgy programs of the 1970s and therapid brill ' of the semiconductor and computerindustrie . relied successfully on the importa-

tion of scientific and technical talent from relatedfields. Many analysts consider the mobility andadaptability of the Nation's scientific and engi-neering work force to be one of its greateststrengths.

Spot shortages among experienced scientistsand engineers with specific areas of expertise, ashave been reported in surveys of electronics andcomputer firms, are probably inevitable. Whena new technology is developed it simply takes timebefore there is a pool of technically trained per-sonnel who are experienced with the new tech-niques. There is no way to speed up this processmore than minimally. The government could pos-sibly help by taking a more active role in promot-ing the retraining of experienced engineers. Mis-sion agency research and development programsin rapidly growing fields could possibly be mod-ified to serve a retraining function. Apart fromsuch assistance, however, the government role inalleviating shortages of scientists and engineersappears quite limited.

OTA finds that the chang'iig college studentdemographics of the coming decade have severalimportant implications for national policy to pro-mote equality of opportunity in science and engi-neering. If the number of college graduates de-clines, it becomes especially important to utilizeall potential human resources to the fullest extentpossible. This means that increasing attentionshould be paid to those groups with historicallyweak participation rates in science and engineer-ing, such as women and some minorities. The in-creasing fraction of college students that will bedrawn from the black and Hispanic populations,which have historically participated in science andengineering education and employment at farlower rates than the white population, imply thatprograms aimed at increasing the participation ofthese two minority groups could be an especiallyimportant source of new talent. Thus, near-termdemographic trends underline the importance ofpromcting equality of access to scientific and engi-neering careers for women and disadvantaged mi-norities.

Two factors stand out as crucial impedimentsto women's participation in science and engineer-ing careers. The first is gender-stereotyped career

7

expectations among younger women entering thescience a_id engineering talent pool. These aremanifested most climatically in the major fieldpreferences of college freshmen where 20 percentof the men rlxvt.yed in 1984 but only 3 percentof the women, lis:ed engineering as their field ofchoice. By contrast, only 4 percent of the men,but 21 per n. it the women, listed education,nursing, or occupational and physical therapy astheir preferred major It appears that fields thatare heavily associated with "men's work" tend tobe avoided by women, just as fields that are stereo-typically associ .ted with women are avoided bymen.

Girls' expectations of ha: they will be able toallocate their tine during adulthood between par-ticipation in the labor force and work in the homealso affects their career decisions. One analystfound that the more girls expect continuous la-bor force participation during adulthood, themore their occupational goals approximate thoseof their male counterparts. She also found thatmale single parents make occupational and laborforce adaptations to parenting that look like theoccupational and labor force plans of youngwomen who expect dual family and work respon-sibilities. As long as women expect to assume themajor role in housekeeping and childrearing, andto sacrifice their professional interests to those oftheir husband, they will be less likely than men,to select occupations like science and engineeringthat require major educational and labor forcecommi,ments.

The second principal factor discouraging womenfrom pursuing scientific and engineering careersis their differential treatment in the work force.Women's attrition rates from scientific and engi-neering careers are 50 percent higher than men'sand their unemployment rates are more than dou-ble. Women's salaries are significantly lower thanmen's in almost all fields of science, in every em-ployment sector and at comparable levels of ex-perience. In academia men are far more likely thanwomen to hold tenure-track positions, to be pro-moted to tenure and to achieve full professorships,even when academic age, field, and quality ofgraduate schoo! attended are controlled for.

The differential treatment of women in thework force is thought by some to be the most seri-

h

ous violation of the principle of equality of op-portunity because it affects people who haveestablished, by virtue of obtaining an advanceddegree, the right to pursue a scientific or engineer-ing career based solely on the quality of theirwork. It also has a significant discouraging effecton female students in the educational pipeline,who see the future benefits of their investment inscience and engineering education eroded by po-tential unemployment and underutilization in thework force. In sum, gender-stereotyped career ex-pectations and differential treatment of womenscientists in the work force are the two major fac-tors discouraging women from entering scienceand engineering.

Among the minority groups that have been his-torically ur 2rrepresented in science and engineer-ing, blacks, Hispanics, and American Indians re-ceive degrees in quantitative fields at less than halfthe rate of whites, reflecting both lower partici-pation in higher education and a lower rate ofselection of quantitative fields. The causes of thisunderrepresentation are not well understood andcould benefit from additional re. arch. The qual-ity of academic preparedness in secondary schoolsis cited by many experts as the greatest factoraffecting tilese minorities' academic performanceand baccalaureate attainment in college. Aca-demic preparedness, however, appears to be re-lated to socioeconomic factors such as parents'educational levels and social class. According toone study, the difference in the choices of quan-titative majors across raci I and ethnic groups areeliminated when the person i: the second genera-tion of a family tc attend college. Social class,according to the same study, sef ms to be a proxyfor a variety of family characteristics that affectschool achievement, incluaing the family's stresson achievement, language models in the home,academic guidance provided by the home, workhabits and activities of the family, and the natureand quality of toys and games available to thechild. They correlate very highly with children'sachievement scores.

Despite the formidable socioeconomic problemsresponsible for the low participation rates inscience and engineering among blacks, Hispanics,and American Indians, there is considerable evi-dence that well-designed intervention programs

8

can assist these groups in obtaining access toscience and engineering careers. To date these pro-grams for minorities have not been rigorously andsystematically evaluated to determine the ingre-dients of success and tailure. Nor has any attemptben made to expand the coverage of the moresuccessful programs. NSF was mandated by Con-gress to take a leadership role in this area, but thusfar it has not done so. However, it appears safeto conclude that the socioeconomic factors thatlead to poor academic performance and an in-ability to remain in the science and engineering"pipeline" among blacks, Hispanics, and Amer-ican Indians can be compensated for, to some de-gree, by well-designed intervention programs.

Finally, OTA examined the increasing partici-pation of foreign nationals in U.S. science andengineering education, which has raised a num-ber of unresolved policy issues. The number offoreign nationals enrolled in American institutionsof higher education has increased by a factor of10 in the past thr.c decades. Foreign students tenonto enroll proportionately more often in engineeing and medical sciences, and less in humanitiesand social sciences. They have assumed a verylarge share, 25 to 45 percent, of enrollments anddoctoral degrees in graduate departments of engi-neering, physics, mathematics, and computer sci-ence, at the same time that the enrollment ofAmerican students in those departments declined.

There is concern that because science and engi-neering education is capital-intensive, and becausemany science and engineering graduate studentsare supported by Federal research assistantships,the United States is subsidizing the training of non-U.S. citizens. However, recent analysis indicatesthat foreign degree recipients appear to contrib-ute more than the full cost of their education totheir host institution. In addition there are for-eign policy and goodwill benefits to the UnitedStates from the education of foreign citizens. Italso appears that many graduate engineering de-partments would have had to curtail their researchactivities substantially without foreign graduatestudents.

There is disagreement as to whether individualsadmitted to the United States as students who

complete a graduate degree here should be al-lowed to remain here to accept employment.Many employers say that they cannot fill certainresearch and faculty positions without foreign na-tionals. Some professional societies, on the otherhand, contend that U.S.-educated foreign engi-neers are taking jobs at lower pay in order to re-main in this country, thus driving down salariesand reducing opportunities for U.S. engineers.

Several authors have made reference to a"growing problem" of foreign nationals on engi-neering faculties. They contend that the draw-backs of an increasingly foreign born faculty arelimitations in their communications skills anddifficulties they may have in adjusting their teach-ing and research styles to the expectations ofAmerican culture. Other observers note that for-eign nationals may be unable to work on indus-try or government projects with national securityor economic competitiveness implications. Thevalidity of these assertions has yet to be docu-mented.

The situation with respect to foreign nationalscan best be summarized by saying that althoughforeign nationals represent a large and growingtraction of science and engineering Ph.D.s, thereis considerable disagreement over whether thisconstitutes a problem.

In general, there does not appear to be convinc-ing cause for concern that demographic trends willlead to shortages of scientists and engineers. Thelabor market for technical personnel has a vari-ety of mechanisms for adjusting to potential sup-ply-demand imbalances, and these appear to workreasonably well, without any great need for gov-ernment intervention. However, the decline in thenumbers of college graduates available for scien-tific and engineering careers in the next decademakes it increasingly impnrtant that all potentialresources for the scientific and engineering workforce be utilized to the fullest extent possible. For:his to happen, the problems experienced bywomen, blacks, Hispanics, and other disadvan-taged groups in entering scientific and engineer-ing careers will have to be better understood andaddressed.

Contents

Page1111

1520

Background and Rafional for This StudyThe Science and Engineering Work ForceDemographic Trends and the Possibility of Futere Shortvick.Findings

List of FiguresFigure No1-1. Population of J.S. Scientists and Engineers by ROA 1063'1-2. Employed Scientists and Engineers by Highest De, 19831-3. Employed Scientists and Engineers by Sector;1-4. Employed Scientists and Engineers by1-5. U.S. Population by. Major Age Group1-6. Population Distribution by Age,1-7. Population Distribution by Aged1-8. 18-24 Year Olds and 25-34 Year1-9. Science/Engineering Bachelor's and'

1-10. Science/angineering Master's1-11. Science/Engineering Doctorates1-12. Comparison of Junior Faculty

WAi; 1983

Page12131314151617s

Awarded by Reid1919

Awarded 22

401:14010

fa;4 ......

A 9

---1Chapter 1

Overview and Findings

BACKGROUND AND RATIONALE FOR THIS STUDY

The key role of the Federal Government in edu-cating and assuring an adequate supply of scien-tists and engineers has been acknowledged sincethe close of World War II. It was reemphasizedin a series of reports from the President's ScienceAdvisory Committee in the Immediate post-Sputnik (1958 to 1962) era which, according toone analyst,'

. articulated the national need for greater num-bers of scientists and engineers . . , and forstronger federal support for the training of man-power for basic research and for university levelteaching.

The current Administration has reaffirmed theFederal commitment to the education and train-ing of scientists and engineers, stating that:3

. . . we have to make sure that we derive educa-tional and training advantages from federally sup-ported researcl- because all of our expectations

'See Henry A Moe, Chairman ''Repert of the Committee on theDiscovery ai,d Development of Scientific Talent, in '. annevar Bush,ScienceThe Endless Frontier, Office of Scientt is Research and De-velopment (Washington, DC U S Government Pnnting Office, July1945), pp 135-186

'Susan Fallows, 'Federal Support for Graduate Education in theSciences and Engineering, background paper prepared for the AdHcc Committee on Government-University Relationships in Sup-port of Science, National Academy of Sciences, 1983, p 5

'U S Congress, House, 'Testimony of Dr George A Keyworth,Director, Office of Science and Technology Policy, Executive Of-fice of the President,' 1983 Science and Technology Posture Hear-ing With the Director of the Office of Science and Technology Pol-icy, Committee on Science and Technology, 98th Cong , 1st sess ,

Feb 3, 1983 (Washington, DC U S G, ernmenf Printing Office1983), p 15

and opportunities for industrial progress call fora growing supply of skilled technical personnel.

The Task Force on Science Policy of the HouseCommittee on Science and Technology finds that"many of the changes in educational and man-power demands are related it a significant wayto general developments in the rates of birth andretirements." It believes that "these broad, dem-ographic chaages, if prudently used in connectionwith other data, might well provide insights thathelp anticipate future changes in enrollments andrelated developments." To better understandwhether demographic trends could be used to helpanticipate changes in the requirements for edu-cation and training of scientists and engineers, tr.1Committee asked OTA to carry out a stady of"Dem:-..,;raphic Trends and the Scientific and Engi-neering Work Force." As part of that study theTask Force on Science Policy recommended thatOTA examine the effects of the "growing inter-est on the part of women and minorities" in pur-suing scientific careers and the "barriers to suchparticipation and the means for lowering them."5This report is presented as OTA's response to theCommittee's request, and ac its contribution tothe national dialog on thi, important aspect ofFederal science policy.

'LT S Congress, House, An Agenda for a Study of GovernmentScience Policy, report by the Task Force un Science Policy. Com-mittee on Science and Technology, 98th Cong , 2d sess , Serial MM(Washington, DC U S Government Printing Office, 1985), p 30.

'Ibid , p 31

THE SCIENCE AND ENGINEERING WORK FORCE

Scientists and engineers represent 3 percent ofthe national work force, but are considered bymany to be a crucial element in the Nation's ef-forts to improve its economic competitiveness and

national security. Professor Karl Willenbrock ofSouthern Methodist University expressed this sen-timent quite vividly in hearings on "Scientists andEngineers: Supply and Demand" before the Sci-

11

12

ence Policy Task Force of the House Committeeon Science and Technology:6

The nation's economic vigor and quality of life,as well as military security, are strongly depend-ent on the number and quality of the engineersand scientists which the U.S. has available bothnow and in the future. Thus, the health and well-being of the system which educates Americanyouth in engineering and science, and enables thepracticing engineer and scientist to stay at theforefront of rapidly developing fields of scienceand technology is a crucial part of the nation'sscience policy.

'U S. Congress, House, "F. Karl Willenbrock, Testimony," Com-mittee on Science and Technology, Science Policy Task Force, July10, 1Q85, p 1.

Employed scientists and engineers numbered 3.5million in 1983, an increase of 50 percent from1976. Dui ing that sa me period the national em-ployed labor force only increased by 11 pe. cent.The scientific and engineering work force is quiteheterogeneous in its composition. Fifty-six per-cent of the total are engineers, predominantly elec-.rical and electronic engineers (13 percent); me-chanical engineers (11 percent); and civil engineers(8 percent). The remaining 44 percent are scien-tists, 11 percent each in the life and social sciences;10 percent each in the physical and computer sci-ences; and 3 percent in the mathematical sciences.Figure 1-1 presents the breakdown of the scienceand engineering work force by field. Among thescientists, one in five has completed the full doc-toral training that is generally considered neces-

Figure 1 -1 .Population of U.S. Scientists and Engineers by Fisk', 1983

Mathematicalscientists

3%

Aeronautical/astronautical

2%

Scientists440/0 Engineers

56%

SOURCE National Science Foundation Science and Technology Resources, NSF-85 305 January 1985 pp 53-65

21

13

sary for a research career. Nearly half hold onlya bachelor's degree and a third have been edu-cated to the master's level. Engineers are evenmore heavily concentrated at the bachelor's andmaster's h vel, as can be seen in figure 1-2.

More than 80 percent of the Nation's engineersare employed in business and industry, with 12percent in government and only 3 percent in edu-cation. The scientists are somewhat more evenlydistributed among the three employment catego-ries, with 51 percent in business and industry, 17percent in government, and 24 percent in educa-timal institutions (see figure 1-3 for more details).

Eighteen percent of the scientists, but only 4 per-cent of the engineers, list research as their primarywork activity. Thirteen percent of the scientists,but only 2 percent of the engineers, list teachingas their primary activity. Development, by con-trast, is the principal work activity of 30 percentof the engineers but only 9 percent of the scien-tists (see figure 1 -4).'

The racial, sexual, and ethnic composition ofthe scientific and engineering work force is alb()

'Science and Technology Resources, NSF 85-305 (Washington,DC National Science Foundation, January 1985), pp 53-65.

Figure 1-2.Employed Szientists and Engineers by Highest.Degree, 1983Scientists/engineerstotal = 3.5 million

12% Other

52% Bachelor's

25% Masters

11% Doctorate

Scientists = 1.5 million Engineers = 1.9 million

1% Other

46% - Bachelor's

20% Doctorate

33% Master's

8% Other

3% Doctorate

58% Bachelor's

21% Master s

SOURCE Naticnal Science Foundation, Science and Technology Resources, NSF 85-305 January 1985, pp 53-65

Figure 1.3. Employed Scientists and Engineers by Sector, 1983

Scientists/engineerstotal = 3.5 million

7% Other

Scientists = 1.5 million Engineers = 1.9 million

12% Academia 8% Othe.

67% Business/Industry

14% Government

24% Academia 3% Academia

51% Business/industry

17% Governmen

5% Other

80% Buslne^ dustry

12% Government

SOURCE National Science Foundation, Science and Technology Resources, NSF 85-305, January 1985, pp 53-85

4. 4

14

Figure 1.4. Employed Scientistsa and Engineers by Primary Work Activity, 1983

Scientists and engineers3,465,900

Scientists1,525,900

Engineers1,940,000

Development9%

alncludes social scientists and esychologists

SOURCE National Science Foundation Sc,ence and Technology Resources NSF 85-305 January 1985 pp 53-65

quite complex and varied. Blacks, who constitute10.4 percent of the national labor force, only con-stitute 2.6 percent of the scientists and engineers,ranging from a high of 6 percent of the socialscientists to a low of 0.6 percent of the environ-mental scientists. Hispanics, who are 5.5 percentof the total labor force, represent only 2.3 per-cent of the Nation's scientists and engineers. Atthe other extreme, Asians, who constitute 1.6 per-cent of the national labor force, make up 4.5 per-cent of the science and engineering work force.

Women make up 43.5 percent of the civilianlabor force, but only 13.1 percent of the scis.a-

tiFis and engineers. They are especially well repre-sented in the mathematical sciences (45.8 percent)and psychology (40.9 percent), and especiallypoorly represented in engineering (3.5 percent)and physical science (11.7).8 There is concern thatthe low participation rates of women, blacks, andHispanics in science and engineering reflect notonly tastes and preferences but also social and cul-tural barriers to equal opportunity in those fields.

"Professional I'Vomen and lfinorities 5th ed (Washington, DCScientific Manpower Commis,Ion, August 1c84), p Q6 , StatisticalAbstract of the United States, 1085 (Washington, DC U S Depart-ment of Commerce, Bureau ot the Census, 1Q85), pp 391, 302

15

DEMOGRAPHIC TRENDS AND THE POSSIBILITYOF FUTURE SHORTAGES

The U.S. population grew from 152 to 234.5million between 1930 a .1c1 1983, and is projectedto increase to 283 million by the year 2010.9 Therate of growth has slowed dramatically, from 19percent between 1950 and 1960 to 11 percent be-tween 1970 and 1980, and a projected 5.6 percentbetween 2000 and 2010. Within that overallgrowth, different subpopulations have experi-enced very different patterns of change, as canbe seen from figure 1-5.10 The school age and pre-school populations (from birth to 17) grew rap-idly between 1950 and 1970, and have essentiallyleveled off since. The college and university agepopulation (18 to 34) underwent significant ex-

Projections of the Population of the United States, by Age, Sexand Race: 1983 to 2080, Series P-25, No 952 (Washington, DCU.S. Department of Commerce, Bureau of the Census, May 1984),table E, p. 7 See also, Statistical Abstract of the United States, 1985,op. cit., table 27, p 26

"Projections of the Population, op cit , table E, p. 7

Figure 1.5.U.S. Population by Major Age Group,1950-2010

340

320

300

280

261-240

220

200

+56%

+ 11 4 %

180

160 + 19%

140

120

100

80

60

40

20

o1950 1960 1970 1980 1990 2000 2010

P P P

>b5

35-64

18-34

<17

YearSOU': Protections of the Population of the United Stales, by Age, Sex and

Race 1983 to 2080, series P 25, No 952 (Washington, DC U SDepartment of Commerce, Bureau of the Census, May 1984) table E,p 7

pansion between 1960 and 1980, will contract insize again between 1990 and 2000, and then ex-pand somewhat between 2000 and 2010. The sen-ior working age population, 35 to 64 years old,will expand dramatically in size between 1980 and2010, as will the total number of elderly people,aged 63 and over.

Figures 1-6 and 1-7" display these populationchanges in somewhat greater detail. They showthe percentage of the population in each age groupin every fifth year from 1950 to 2000 and the year2010. By following the bars from left to right oneach graph one can trace the effects of the "babyboom" of the post World War II era on each pop-ulation group. The successive peaks in popula-C 3n-1960-70 for the 5 to 13 year olds; 1975-85for the 18 to 24 year olds; 1990-2000 for the 35to 44 year oldsare exhibited, as are the subse-quent dramatic declines in the percentage sharesof the different groups following the "babyboom," and the rather modest "baby boom echo"of the early 1980s.

The groups of concern in this essay are the 18to 24 year olds and the 25 to 34 year olds, thetwo subpopulations that contribute in greatestnumbers to undergraduate and graduate enroll-ments. Therefore, it is worth looking at the dem-ographic trends for these two subgroups in some-what greater depth. Figure 1-8" presents thepopulation trends for 18 to 24 -.1-, d 25 to 34 yearolds over the 1950-2010 time period. The figureshows the dramatic increase in the former groupbetween 1955 and 1980, doubling from 15 to 30million in that period. It shows as well the 23-percent decline -irojected for the 1980-95 timeperiod, followed by an 18-percent increase be-tween 1995 and 2010. For the 25 to 34 year oldgroup, there is a doubling in the 1935-90 timeframe followed la, a decline of 16 percent between1990 and 2000.

Each year about 300,000 college seniors, or 30percent of the graduating class, receive theirbachelor's degree in a science or engineering field.

"lb.d., table F, p 8"Ind , table E, p 7

4. 1

16

Figure 1-6.Population Distribution by Age, 1950-201020

15

15 10

5

0

5 10

2

Under 5 III Ages 5-13

1 11950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2010

Year

Ages 14-17 Ages 18-24

I 1 1 1 I1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2010

Year

SOURCE Projections of the Populatr,n of the United ',tales by Age Sex and Race 1983 to 2080 series P 25 No 952 tWashingtott DC U S Department ofCommerce, Bureau o' e Census, May 1984) table F, p 8

Figure 1.7.Population Distribution by Age, 1950-2010

IIIIAges 25-34 El Ages 35-44

17

30

25

20

15

10

5

0

'950 1955 1960 1965 1970 1975

Year

1580 1°85

IIIIAges 45-64 [1:1 65 and over

1990 1995 2000 2010

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2010

YearSOURCE Promotions of the Population of the United Stales, by Age, Sex and Race 1983 to 2080 series P 25 No 952 (Washington DC U S Department of

Commerce Bureau of the Census May 1984), table F, p 8

-, ,4.0

18

Figure 1.8.-18-24 Year Olds and 25-34 Year Oldsin the U.S. Population, 1950-2010

45,000 .40,000

35,000

30,00064

25,000

20,000 I"J

15,000

10,000

.

5,000

1950 1980 1970 1980 1990

Year

2000 2010

NOTE Flgv3s for 1983 to 2010 are projectedSOURCE Projections of the Population of the United States, by Age, Sex and

R. -e 1983 to 208C. series P 25, No 952 (Washington DC U SC partment of Commerce, Bureau of the Census, May 1984), table E,p7

This overall level has remained nearly constantfor close to a decade, but there hax,e been dra-matic fluctuations amo.ig fields (see examples onp. 21).13 Science and engineering masters degreeshave remained relatively constant in the 53,500to 56,500 per year range since 1972, with socialscience and engineering each accounting for morethan 25 percent of the total. The number of scienceand engineering doctorates peaked at 19,000 in1972 and has fluctuated between 17,000 and18,000 a year since 1976.14 (See figures 1-9, 1-10and 1-11.15)

The relatively constant rate of production ofscientists and engineers over the past decade corldbe significantly affected by the projected dee..nein the cc" .!ge age population in the near future.Labor market specialists estimate that the decreasein 18 to 24 year- olds discussed above could leadto a drop in college enrollments of 12 to 16 per-t. It between now and 1995.16

"Science and Technology Resources, op cit , pp 73-74"Science and Engineering Personnel A National Overview, NSF

85-302 (Washington, DC National Science Foundation, 1985), pp154-156

"National Patterns of Science and Technology Resources, 1;84,NSF 84-311 (Wasb:ngton, DC National Science Foundation, 1984),pp 29-30

Research Excellence Through the Year 2000 (Washington, DCNational Research Council, Commission of Human Resources, 1979),pp 12-15, William Bowen, Report of the President (Princeton, NJPrinceton University, April 1981), pp 71-15

Figure 1-9.SciencelEngineering Bachelor's and FirstProfessional Degrees Awarded by FieldNumber As a percent of all bachelors and firetprofeeelonat degrees

Thousands Percent320 100

280

240

200

180

120

to

40

C

7A

30

20

15

10

0 01960 82 64 86 68 TO 72 74 76 78 60 82 1960 62 64 66 68 70 77 74 76 78 80 82

Year Year

f.

4

aInCIVtlH payChology'Includes compute, sclenCecIncludes environmental whines

SOURCE National Science Foundation, National Patterns of Science and Technology Resources NSF 84-311 1'84 pp 29 and 30

60

55

50

45

40

4035vi 30

125

20

15

10

5

01960

19

Figure 1 -10.SciencelEngineering Master's Degrees by Field

Number

82 84 68 88

°Includes psychologybincludes computer sClenceCincludea environmental science

SOURCE National Science 'oundatIon, herons, Patterns of Science and TechnologyResources, NSF 84-311, 1984, pp 29 and 30

70 72

Year

74 78 78 80

100

30

2

120

1 15

82

10

5

0

As a percent of all master's degrees

1960 82 64 es 68 70 72

Year

Figure 1 11.SciencelEngineering Doctorates Awarded by Field

Number

20

15

0vc21

1-

5

100

70

60

50

E8 40k

30

0.4.

1960 82 64 66 68 70 72 74 78 78 80 82 83Year

'Includes psyCnologyb Includes computer science%dudes environmental science

10

01980 62 64 66

74 78

As a percent of all doctoral degrees

78 80 82

68 70 72Year

SOURCE National Science Foundation National Patterns of Science and Technology Resources NSF 84-311 1984 PP 29 an J JO

74 78 78 80 82 83

1

20

In addition to these changes in the overall num-bers, there will be qualitative changes in the stu-dent mix that could affect science and engineer-ing. The percentage of minorities in the 18 to 24year old cohort will increase from 20 to about 27percent by 1998," so the increasing participationof blacks, Hispanics, Asian Americans, and otherminorities in higher education that took place inthe 1970s (minority enrollments increased 86 per-cent from 1972 to 1982) can be expected to con-tinue. (Thc decline in the rate of college enroll-ment among black high school graduates since1978 casts some doubt on this prediction as it re-lates to the black community.) These groups, withthe exception of Asian Americans, have histori-cally participated less activel) in science and engi-neering education than the majority population.The numbers of women, part-time students andstudents in 2-year institutions all increased byabout 70 percent in the decade from 1972 to 1982,approximately twice as rapidly as the overall in-crease in higher education enrollments." Thesegroups also participate in science and engineer-ing education at lower rates than their white malefull-time counterparts at 4-year colleges and uni-versities.

Assuming that science and engineering degreescontinue to represent no nacre than 29 to 30 per-cent of the total baccalaureates awarded in anyyear, as they have since 1972, the decrease in the18 to 24 year old population could lead to a de-cline in the number of bacheior's level scientistsand engineers being produced by the Nation's

"Statistical Abstract, 1985 op at , pp 26-34"Digest of Education Statistics, 1983-1984 (Washington, DC Na-

tional Center for Education Statistics, 1984), p 98

higher education system of a corresponding 15percent by 1995. Similarly, the decline in 25 to34 year olds projected for the 1990-2000 timeframe could lead to a possible decrease in the num-ber of science and engineering doctorates pro-duced in that time period. (The master's degreehas considerably less meaning in science and engi-neering than it does in other fields, such as busi-ness, social work, and education, so it will notbe discussed at any length in this technical memo-randum.)

To investigate the likelihood of this possible de-cline taking place, and its possible consequencesfor the Nation's scientific and technical enterprise,OTA examined the historical record on the rela-tionship between science and engineering degreesand demographic trends and on the labor mar-ket for science and engineering baccalaureates andPh.D.s in industry and academia. OTA also re-viewed recent projections of supply and demandfor scientists and engineers, consulted extensivelywith scientific and engineering labor marketspecialists and carried out a number of irp.I.epend-ent studies and simulations on its own. OTA'sfindings from its investigation of science and engi-neering personnel issues are Presented below.Findings 1 to 3 deal with population trends andthe education and employment of science andengineering baccalaureates and Ph.D.s. Findings4 to 6 deal with projections of supply and demand,

nd the functioning of the labor market for tech-nical personnel. Findings 7 to 10 examine the bar-riers to equality of opportunity in science andengineering careers experienced by women andsome racial and ethnic minorities, and the spe-cial problems of foreign nationals in science andengineering education and employment.

FINDINGS

in a science or engineering field. Less than 0.5 per-cent achieve the doctorate in one of those fields."These very small pei ..zntages imply that changesin the overall size of the age cohort should havefar less of an effect on the number of scientistsand engineers produced in a given year than var-

Finding 1

Population trends, although important, do nothave as great an impact on the supply of scien-tists and engineers as do career choices and mar-ket forces.

Less than 7 percent of the members of a givenage cohort currently receive a bachelor's degree

"Science and Technology Resources, op cit , pp 74 and 78, Pro-jections of the Population, op cit , table 5, p 3ta

- --,

4: d

iations in the rate at which students choose to en-ter those fields. A slight increase in the rate ofselection of science and 7.ginecring careers amongcollege age students could more tnan compensatefor the decline in 18 to 24 year olus projer.ed forthe 1982-95 time frame. For example, if the rateof attainment of science and engineering bache-lor's degrees were to increase from the 1982 levelof 6.3 percent of the 22 year old population tothe 1973 level of 8.2 percent,2° that would causea rise in the number of degrees awarded in thesefields of 20 percent.

Less than two-thirds of those receiving scienceand engineering baccalaureates in recent yearshave actually become part of the science and engi-neering work force. The National Science Foun-dation (NSF) conducted a survey of science andengineering bachelor's degree holders in 1982 todetermine the experience of the lasses of 1980 and1981 in making the transition from school towork. It found that only 43 percent of the scienceand engineering baccalaureates from those 2 yearswere employed in a science or engineering fielda year or two later." An additional 21 percentwere enrolled in graduate school full time, al-though not necessarily in science or engineering.Twenty-eight percent were employed outside ofscience and engineering, 5 percent were unem-ployed and seeking employment, and 4 percentwere outside the labor force. Thus, less than two-thirds of the 1980-81 science and engineering bac-calaureates were actuagy in the science and engi-neering work force a year after graduation. Thispercentage varied from a low of 43 percent in thesocial sciences to a high of over 80 percent in engi-neering, computer science, and the physical sci-ences, with -nathematics and the life sciencesdirectly it the middle, at 70 and 67 percent, re-spectively.

There is, moreover, no national market forscientists and engineers as a group. Rather, thereare individual markets for graduates trained inparticular disciplines, and these markets and dis-ciplines experience very different conditions at theJame time. For example, there is currently alleged

"Science and Engmeenng Degrees 1950-1980, NSF 82-307 (Wash-irgton, DC National Science Foundation, 1982), 60

"Nabonal Patterns o4 Science 6r Technology Resources, 1984, opcit , p 24

21

to be an 8 percent vacancy rate in engineeringfaculty positions zt a time when new Ph.D.s insociology, mathematics, and other tields are hav-ing trouble finding university employment.22 NSFprojects that the demand for electrical and aero-nautical engineers and for computer specialistscould exceed the supply of graduates in thosefields by as much as 30 percent over the 1981-87time frame, while at the same time there will besignificantly fewer openings for biologists, chem-ists, geologists, physicists, mathematicians, chem-ical, civil, and mechanical engineers than therewill be trained degree holders." Nearly 28 per-cent of the science baccalaureate holders in 1981obtained employment outside of science and engi-neering in the same year that 40 percent of thefirms responding to an NSF survey reported short-ages of engineers.24 Thus, it is individual dis-ciplines, especially those linked with high growthor defense-oriented industries, which could experi-ence problems; not "science and engineering" asa whole.

This finding is encouraging, because individ-ual disciplines have shown great ability to growand shrink as market conditions require. Thenumber of engineering baccalaureates, for exam-ple, more than doubled between 1976 and 1984."The number of bachelor's degrees in computer andinformational science tripled between 1977 and1982, At the same time those in education, for-eign languages, anthropology, history, and sociol-ogy all fell by 30 percent or more." Thus, col-lege students appear to be highly responsive tomarket signals, and appear tc., shift their careerchu, ..-!s dramatically toward fields that promisegreater occupational rewards.27 Engineering andcomputer science have b en highly marketable

"Michael S McPherson, "The State of Academic Labor Markets,"November 1984, draft, p 5 To be published in B L R. Smith (ed ),The State of Graduate Education (Washington, DC The BrookingsInstitution, 1985)

"Projected Response of the Science, Engineering and TechnicalLabor Market to Defense and Nondefense Needs 1982-1987, NSF84-204 (Washington, DC National Science Foundation, 1984), p. 43.

"National Patterns of Science and Technology Resources, 1984,op cit , p 24

"Engineering Education and Practice in the United States (Wash-ington, DC National Research Council, 1985), p Si, and ManpowerComments (Washington, DC Scientific Manpower Commission,January-February 1985), p 34

"McPherson, op cit , table 1"Sue E Berryman, in an exhaustive study of "The Adjustments

of Youth and Educational Institutions to Technology Generated

22

majors over the past 5 to 10 years, as measuredby the numbers of job offers and level of startingsalaries for bachelor's degrees in those fields. For-eign languages, history, sociology, and anthro-pology, by contrast, are all fields that have ex-perienced relatively weak job markets in the pastdecade.

Finding 2

Levels of graduate enrollment and Ph.D. pro-duction in science and engineering appear to belargely independent of general demographictrends.

Between 1960 and 1970 the number of full-timegraduate science and engineering students enrolledin U.S. institutions of higher education increased150 percent, from 78,000 to 188,000." The num-ber of science and engineering Ph.D.s awardedincreased 183 percent, from 6,300 in 1960 to17,700 in 1970.29 At the same time the popula-tion of 22 to 34 year oldsthe group that sup-plies the majority of graduate students increasedonly 18 percent."

Between 1970 and 1982 the situation changeddramatically. Graduate enrollments in science andengineceing increased by only 36 percent; annualdoctorates awarded decreased by 1 percent, butthe graduate school age population increased bymore than 50 percent. Thus, there is no evidenceof a direct relationship between graduate enroll-ments and Ph.D.s in science and engineering, andchanges in the graduate school age population.

The dramatic expansion in Ph.D. productionand graduate enrollment in science and engineer-ing that occurred in the 1960s was apparently re-lated to four simultaneous developments. First,Federal support for graduate education in general,and science and engineering education in particu-lar, grew enormously in that decade. The num-

Changes in Skill Requirements," Research Report Series, RR-85-08(Washington, DC National Commission for Employment Policy.May 1985), finds that "young people generally are responsive tochanging market signals, they alter their fields of study, increaseor decrease the amount of time they spend on formal education,and are very mobile both geographically and occupationally' (p I )

"Fallows, op cit , p. 10"Ibid , p 11"Statistical Abstract of the United States, 1979 (Washington DC

U.S. Department of Commerce, 3ureau of tie Census, 1979), pp160 and 29

ber of Feueral graduate fellowships and trainee-ships increased from 8,000, of which 5,000 werein science and engineering in 1960, to 52,000, ofwhich 36,000 were in science and engineering in1969. The number of research assistantships inscience and engineering increased from about5,000 in 1954 to over 15 000 in 1969.'1 At the sametime Federal support was expanding, State andinstitutional aid to graduate education, especiallyin the form ,f teaching assistantships, also in-creased dramatically, to more than 50,000 grad-uate students in 1969.32

In addition to increased Federal and State sup-port for graduate education there was a third fac-tor, a surging demand for Ph.D.s, especially inacademia, in that e!-a. The number of juniorfaculty opening, exceeded the number of doc-torates awarded by 50 percent (260,000 openings;170,000 doctorates) between 1961 and 1970." (Seefigure 1-12.34) In addition, the Federal research

"Federal Inter-agency Committee on Education, "Report on Fed-eral Predoctoral Student Support," April 1970, tables 1-10 R GSnyder, 'The Effectiveness of Federal Graduate Education Policyand Programs in Promoting an Adequate Supply of Scientific Per-sonnel," contractor report prepared for :he Office of Te..hiologyAssessment, June 1985, p lz

"Snyder, op cit , p 18"Allan Cartter, Ph D s ana he Academic Labor Market (New

York McGraw-Hill, 1976), p .01"Ibid , p 129

Figure 1.12.Comparison of Junior FactetyOpenings With Earned Doctorates Awarded

(actual 1948.73)

1950 1960

YearSOURCE Allan Cartter PhD s and the Academic Labor Market New 'York

McGraw Hill, 1976) o 129

1970

.... ..

Al 1

23

and development (R&D) budget quadrupled inthat decade, creating a sizable demand for doc-toral scientists and engineers as researchers.

Finally, popular opposition to the Vietnam Warled many male college graduates to enroll in grad"ate school to avoid the draft.

In the 1970s, when the growth in academic de-mand and Federal R&D funding slowed, the mil-itary draft ended, and the level of Federal sup-port for graduate fellowships and traineeshipsdecreased dramatically, Ph.D. production beganto fall. It fell most rapidly in those fields whereFederal fellowships and traineeships and R&Dfunds declined most severely (i.e., engineering,mathematical sciences, and physical sciences).

Finding 3

A combination of faculty and student demo-graphics will lead to a weak academic market fornew Ph.D.s over the next decade, followed by anupsurge in academic hiring between 1995 and2010.

The projected decline in student enrollmentsdiscussed above, coupled with a low retirementrate among current faculty (most of whom werehired in the late 1960s and will not retire until thelate 1990s), will lead to a very weak academicmarket for new Ph.D.s in the next decade. Thisdecline was predicted by a number of scholars inthe 1970s, and has been confirmed by analysescarried out this year by OTA and the NationalAcademy of Sciences." According to the OTAanalysis, the total academic demand for newscience and engineering doctoral degree recipientsshould average about 3,000 tc 6,000 per year from1983 to 1998. This is less than one-third the an-nual number of new Ph.D.s projected for theperiod under the assumption that such degrees willremain constant at about 18,000 per year, and 40to 80 percent of the rate of hiring for new scienceand engineering Ph.D.s in academia that existedin the 1978-83 time frame (about 8,000 new hiresper year). After 1998 the demand for new scienceand engineering faculty and research personnelin academia should increase substantially, due toincreases in both enrollments and retirements.

"National Researrh Council, Forecasting Demand for UniversityScientists and Engineers Proceedings of a Workshop (Washington,DC National Academy of Sciences, 1985), p 1

According to OTA analysis, the academic de-mand for new science and engineering Ph.D.scould be roughly the same from 1998 to 2013 asthe level of hiring that took place during the 1970s.The pattern of decline followed by incr.( se willbe experienced separately over the next quartercentury by each of the major science and engi-neering fields. (See chapter 3 for the analysis thatsupports these projections.)

The trends discussed above pose two significantpolicy-related questions. First, should steps betaken now to increase the demand for new Ph.D.sduring the decline decade of 1988 to 1998, in or-der to smooth out the hiring pattern over the nextquarter century and prevent potential graduatestudents from becoming discouraged from pur-suing a research career due to the weak near termacademic market? The National Institutes ofHealth (NIH) system for training and supportingbiomedical research personnel appears to be anexample of a self-conscious and explicit Federalpersonnel policy in an important scientific fieldthat could be emulated elsewhere. Consistent Fed-eral support for biomedical research it. terms offunding of research and training has producedboth a substantial supply and demand for inves-tigators. This has created a steady-state systemin Arhich there have been no major shortages ofbiomedical personnel, although some analysts be-lieve there to be a surplus of biomedical Ph.D.stoday. However, it is a large system which wouldbe difficult to reduce in significant ways withoutmajor dislocation;.

Unlike most other fields of science, there is aninstitutional feedback mechanism between NIHand the Institute of Medicine (IOM) that ensuresa periodic assessment of the status of biomedicalresearch personnel. This represents, to some ex-tent, informed decisionmaking in the implemen-tation of research training programs. The currentnumber of trainees is quite close to the numberrecommended by the IOM committee. Allocationof Federal support to either predoctoral trainingor postdoctoral training is a method of fine tun-ing that can potentially be used to control the sup-ply of biomedical investigators. (See appendix Afor an in-depth analysis of Federal education andmanpower policies with respect to biomedical per-sonnel.)

24

Second, in the expansion period of 1998-2013,will the combined requirements of industry andacademia for science and engineering Ph.D.s ex-ceed the annual supply? The number of scienceand engineering Ph.D.s employed in industry hasgrown at the rate of R percent per year over thepast decade.36 If this rate of growth continues, theindustrial demand for new science and engineer-ing Ph.D.s could exceed 10,000 per year by thefirst decade of the 21st century. The combinedacademic and in *ustrial demand for new scienceand engineering Ph.P.s could then exceed the levelof Ph.D. production for that period.

Finding 4

Long-term projections of supply and dem?..idfor scientists and engineers in the economy as awhole are inherently unreliable.

Apart from academia, where known demo-graphic trends exert considerable influence on thedemand for new faculty, long-term projections ofdemand for scientists and engineers in the over-all economy are fraught with uncertainty and arerelatively unreliable. Such projections depend onassumptions about the future behavior of varia-bles such as GNP growth, defense spending, tech-nological change, and Federal and industrial R&Dexpenditures, which themselves are not knownwith any degree of certainty. Unexpected changesin any of the input variables can cause major in-accuracies in the projections. For example, the Bu-reau of Labor Statistics (BLS) projects the require-ments for different occupations 10 .nd 15 yearsinto the future, and provides an "early warning"to high school and college students through its an-nual Occupational Outlook Handbooks, whichare distributed to guidance counselors. Laboreconomist Lee Hansen took the projections madeby BLS in 1960 and 1965 for the number of engi-neers that would be required in different dis-ciplines in 1975 and 1980, and compared themwith the actual numbers that were employed inthose years." He found that BLS overprojectedthe requirements for aeronautical, civil, and me-chanical engineers by 20 to 55 percent. The er-rors were caused by the understandable failure of

"Characteristics of Doctoral Scientists and Engineers in the UnitedStates, NSF 85-303 (Washington, DC National Science Foundation,1985), table 2, pp yin and ix.

"Labor Market Conditions for Engineers Is There a Shortage?(Washington, DC: National Research Council, 1984), pp 89-92

the BLS to anticipate the cutback in Federal R&Dexpenditures in the early 1970s and the recessionin the mid-1970s. However, the BLS overprojec-tions would have lead analysts to overestimatethe growth component of the annual demand forengineers between the early 1960s and the late1970s by 120 to 600 percent.

Projections of supply are equally problematic.It is extremely difficult to predict the career choicesof undergraduates or graduate students, or theirresponsiveness to changes in demand in techni-cal fields. In the absence of such a predictive ca-pability, analysts in the past tended to assume thecontinuation of existing trends, an assumptionwhich can lead to gross inaccuracies. For exam-ple, in 1967 NSF, extrapolating from existingtrends, projected there would be 30,500 doctoratesawarded in science and engineering in 1975.38 Theactual number turned out to be 17,784. A largenumber of scientific "manpower" analysts in thelate 1960s and early 1970s made similar errors forthe same reasons. In the absence of a causal modelthat can accurately predict changes in career selec-tion patterns, any supply projections are no bet-ter than educated guesses.

Given the problems of projecting supply anddemand, predictions of shortages or surplusesbased on such projections are not likely to be veryreliable. This is not, however, a serious problem,for reasons that are discussed in the next twofindings.

Finding 5

Labor markets adjust to supply- demand gaps.

Most projections of shortages assume tacitlythat demand and supply are independent, so thatthere can be no adjustment to projected supply-demand gaps. This assumption makes sense, how-ever, only under certain very restrictive con-ditions:

1. that demand for the final product is rela-tively unaffected by labor costs;

2. that supply is not appreciably affected bywage changes; and

3. that the skill shortages are unique, in thatworkers possessing them cannot be replaced

'"The Prospective Manposser Situation for Science and EngmeenngStatf in Universities and Colleges 19o5-1075 (Washington, PC Na-tional Science Foundation, 19671, p lb

:i 3

1

by workers from other occupations or bynew technology.

If any of these conditions are violated, then a pro-jected supply-demand gap can be closed by mar-ket adjustn tent. These conditions are very restric-tive: They apply to only a few limited sectors,

-N as some parts of the defense industry andt _ ..in startup firms that require highly special-ized personnel.

The essence of a labor market is that it adjusts.A projected gap between supply and demand usu-ally does not indicate an impending shortage, butrather that some sort of adjustment must takeplace. The adjustment can be made either by em-ployees or employers, or both. On the employ-ees' side, an increasing number of entry levelprofessionals can become trained in the shortagearea, or experienced workers from neighboringspecialties can move into the snortage occupation.On the employers' side, they can offer higher sal-aries; increase their search and recruiting efforts;rearrange jobs to utilize the available skills, edu-cation, and experience more efficiently; or makelarger investments in internal training and retrain-ing of their existing work force. Cutting back out-put, or the level of research and development, dueto personnel shortages is probably the responseof last resort.

Labor market and scientific "manpower" spe-cialists consulted by OTA agreed that it is moreimportant to try to understand the process of ad-justment of the technical labor market to supply-demand imbalances than it is to make long-termprojections. There can be significi_lt problemswith labor market adjustments that need to be un-derstood by polizymakers. For example, althoughstudents shift their career interests to follow sig-nals from the marketplace, as we have seen above,this response is not a short-term remedy to asupply-demand gap because it takes 4 years totrain a new engineer and 6 years (from the bac-calaureate) to train a Ph.D. scientist. In a fast-moving market, the needs of employers !naychange by the time entering students graduate.This lag effect is especially powerful for occupa-tions that are subject to sudden unanticipatedsurges in demand, as in the semiconductor andcomputer industries in recent years. In addition,universities may not be able to provide sufficientresources to meet new demands on their curric-

25

ula, as has recently been the case in engineering.Both universities and students may misjudge thedirection of the marketplace, leading to furthermisallocation of resources.

Occupational mobility from related fields is theshort-term response to a shortage. The primaryconcern about occupational mobility from the em-ployer's point of view is that it may lower thequality of work or require expensive retraining.From a societal point of view, however, occupa-tional mobility appears to have served the Na-tion well in meeting its needs for technical per-sonnel. The Manhattan project in World War 11,the Apollo program in the 1960s, the environ-mental and energy programs of the 1970s, and therapid buildup of the semiconductor and computerindustries all relied successfully on the importa-tion of scientific and technical talent from relatedfields. Many analysts consider the mobility andadaptability of the Nation's scientific and engi-neering work force to be one of its greateststrengths.

Employer adjustments, such as extensive searchesand retraining of personnel from related fields,involve considerable costs. For example, one ana-lyst estimates that the cost of recruiting a newengineer can run as high as 1 year's salary. Rear-rangement of jobs within the firm can interferewith the distribution of incentives via rank andpromotion. Despite these costs, several of themore mature computer companies, such as DataGeneral Corp. and Digital Equipment Corp., haveadopted systematic training approaches. Facedwith a continuing series of shortages, they haveshifted from their traditional "buy" to a "make"strategy for obtaining skilled personnel. This helpslower the turnover rate and increases the supplyof experienced engineers, who are in short supply.

Finding 6

The Federal role in alleviating potential short-ages of technical personnel appears limited toassistance for education and retraining.

OTA carried out an analysis of the labor mar-ket in electrical engineering, a profession that hasexperienced reported shortages in the recent past,is heavily involved in areas of rapid technologi-cal change, and faces demand from both the mil-itary and civilian sectors of the economy. OTA

26

found that the supply of entry level electrical engi-neers is likely to stay nearly in balance with de-mand for the f "reseeable future, as occupationalmobility makes up the difference between thenumber of entry level engineers needed and thenumber of college graduates actually holding elec-trical engineering degrees. However, some educa-tional institutions report that they do not havesufficient resources to train all the potential elec-trical engineering students who are interested.

The spot shortages of experienced engineerswith specific areas of expertise reported in sur-veys of electronics and computer firms in the re-cent past will undoubtedly continue. When a par-ticular technology gets "hot," engineers who knowthe field and can Tlandle the overall design of aproject are in great demand. If the technology isnew or has not been used extensively in the past,the simply of engineers who combine these attrib-utes can fall short of demand. Unfortunately,these shortages are not of a type amenable to gov-ernment intervention. The government might beable to help by promoting the retraining of ex-perienced engineers into new startup fields. Mis-sion-agency sponsored R&D programs could con-ceivably serve that retraining function. TheFederal energy and environment programs of the1970s, for example, helped retrain significantnumbers of engineers to become specialists inthose two fields.

As to the effects of increased military spend-ing on the market for electrical engineers, OTAdid not find this to be a significant cause for con-cern. The most definitive short-term projectionsby NSF" indicate that a shift from low to highdefense spending would only increase the demandfor electrical engineers by 4 to 5 percent, turninga slight shortage into a slightly larger or- In asurvey of firms that employ electrical engineersin the Boston area, OTA found that shortages arenot perceived to vary much by the level of mili-tary procurements. Many respondents felt that thetwo climates of military and civilian work are sovery different that there is low mobility betweenthe two "sectors."

"Projected Response of the Science, Engmeenng ano -echmcalLabor Market to Defense and Nondefense Needs 1982-1987, opcit., p. 43.

Finding 7

Near-term demographic trends enhance the im-portance of promoting equality of access to scien-tific and engineering careers for women and dis-advantaged minorities.

The Science Policy Task Force recommendedthat the study of demographics and the scienceand engineering work force also examine issuesrelated to the participation of women and mint _-ities in science and engineering careers, with spe-cial attention paid to the identification of barriersto participation and means of lowering those bar-riers." OTA finds that the changing college stu-dent demographics of the coming decade haveseveral important implications for national pol-icy to promote equality of opportunity in scienceand engineering. Since the number of college grad-uates will decline substantially, it becomes espe-cially important to utilize all potential human re-sources to the fullest extent possible. This meansthat increasing attention should be paid to thosegroups with historically low participation ratesin science and engineering, such as women andsome minorities. The increasing fraction of col-lege students that will be drawn from the blackand Hispanic populations, which have historicallyparticipated in science and engineering educationand employment at far lower rates than the whitepopulation, imply that programs aimed at increas-ing the participation of these two minority groupscould be an especially important source of newtalent.

Finding 8

.7ender-stereotyped career expectations anddifferential treatment of women scientists in thework force are the two major factors discourag-ing women from entering science and engineering.

The literature on the factors that discouragewomen from participating in science and engineer-ing careers is extensive and well developed. Twofactors stand out as crucial. These are gender-stereotyped career expectations among youngerwomen entering the scicrmt and engineering tal-

"An Agenda for a Study of Coyp 31

. I 0

. ment Science Policy, op at.,

I

27

ent pool, and differential treatment of womenscientists and enginec.s in the work force.

Gender-stereo'yped career expectations aremanifested most dramatically in the major fieldpreferences of college freshmen, where 20 percentof the men, but only 3 percent of the women sur-veyed in 1984 listed engineering as their field ofchoice." By contrast, only 4 percent of the men,but 21 percent of the women, listed education,nursing, or occupational and physical therapy astheir preferred major. In other sciences, womenwere more likely to select a social or biologicalscience major than men, and slightly less likelyto be interested in the physical sciences. Thesedifferences are often explained as the result ofwomen's alleged lower preference for or lesserability in quantitative fields. But the fact thatfreshmen women select mathematics, accounting,and business at the same rate as men casts doubton that explanation. Rather, it appears that fieldsthat are heavily associated with "men's work" tendti be avoided by women, just as fields that are.ereotypically associated with women are avoided

by men.

A second aspect of gender-stereotyped careerplanning has to do with girls' expectations of howthey will be able to allocate their time duringadulthood between participation in the labor forceand work in the home. Sue Berryman finds that"the more that girls expect continuous labor forceparticipation during adulthood, the more their oc-cupational goals approximate those of their malecounterparts." She also finds that "since careergoals seem lo determine educational investments,gender differences in occupational expectationsbecome key to understanding gends differencesin high school mathematics participation,"42 a keyelement in pursuing a science or engineering ca-reer. These expectations are in fact derivative ofgender stereotyping in the work force as a whole.In our society women are still expected to assumethe major role in housekeeping and child rearingand to sacrifice their professional interests to those

"A.W Astir), et al , The American Freshman National Normsfor Fall 1984 (Los A. geles, CA Cooperative institutional ResearchProgram, University of Californ a at Los Angeles, December 1984)pp 16-18, 32-34

"S le E Berryman, -Minorities and Women in Mathematics andScience Who Chooses These Fields and Why? May 1985, type-script, p 14

of their husband. As long as that situation exists,women will be less likely than men to select oc-cupations like science and engineering that requiremajor educational and labor force commitments.Berryman finds, interestingly enough, that malesingle parents "make occupational and labor forceadaptations to parenting that look like the occupa-tional and labor force plans of girls who expectdual family and work responsibilities."43

Women's experience in the science and engi-neering work force is decidedly different from thatof men. Women's attrition rates are 50 percenthigher than men's: in 1982, 25 percent of womenscientists and engineers but only 16 percent of menwere outside of the science and engineering workforce." Women's unemployment rates are signif-icantly higher: 4.5 percent v. 1.9 percent for men.Women's salaries are significantly lows-- thanmen's in almost all fields of science, in every em-ployment sector and at comparable levels of ex-perience, although recent female entrants to thescience and engineering work force are doing con-siderably better than their more senior counter-parts. In academia, men are far more likely thanwomen to hold tenure track positions, to bepromoted to tenure, and to achieve full profes-sorships. This holds true even for men and womenof identical academic age (number of years sincereceiving the doctorate), field, and quality of grad-uate school attended.45

The differential treatment of women in thework force most directly violates the principal ofequality of opportunity because it affects peoplewho tvi..P established, by virtue of obtaining anadvanced degree, the right to pursue a scientificor engineering career based solely on the qualityof their work. It also has a significant effect onfemale students in the educational pipeline. Awoman student in a department where no womenhave been hired or promoted to tenure is not likelyto form a positive picture of her future employ-mer.t prospects. Nor will she experience the kindof role model which the literature on equality of

"Ibid"lbw , p 7'For details see Women and Minorities in Science and Engineer-

ing (Washington, DC National Science Foundation, January 1984);and Career Outcomes in a Matched Sample of Men and WomenPh D s (Washington DC National Research Council, 1981)

28

opportunity suggests is desirable to assist her inidentifying with her future profession. Finally, thehigher unemployment and attrition rates amongwomen in science and engineering as comparedto men represent a serious waste of human re-sources cultivated at considerable expense to theNation and the individual.

Finding 9

Blacks, Hispanics, and American Indians areaffected by a variety of socioeconomic factors thatlead to poor academic performance and an in-ability to remain in the science and engineeringeducational "pipeline."

Blacks, Hispanics, and American Indians re-ceive degrees in quantitative fields at less than halfthe rate of whites, due to their tendency both toparticipate in higher education and to select quan-titative fields at substantially lower rates." Thequality of academic preparedness in secondaryschools is cited by many experts as the greatestfactor affecting these minorities' academic per-formance and baccalaureate attainment in college.Greater attrition levels in science and engineer-ing among minorities correlate very highly withpoor academic preparedness in high school asmeasured by grade averages- and class rank. Prob-lems with academic preparedness, in turn, appearto be related to socioeconomic factors such as par-ents' educational levels and social class.

Parents' education appears to be an especiallyimportant factor. Students whose parents haveobtained college or graduate degrees are moreoften enrolled in quantitative majors than studentswho have less well-educated parents. Sue Berry-man finds that "being second generation collegenot only increases, but also equalizes, the choicesof quantitative majors across white, black, Amer-ican Indian, Chicano, and Puerto Rican collegefreshmen."' A study carried out for the Depart-ments of Defense and Labor in 1980 found thatthe strongcst single predictor of reading abilitiesand scores on the Armed Forces QualificationsTest was the mother's educational level." Youths

"Ber:yman, op cit , pp 2-3"Ibid , p 17"Betty M Vetter, The Emerging DemographicsEffect on Na-

tional Policy in Education and on a Changing Work Force back-ground paper prepared for the Office of Technology Assessment,pp 4-5

whose mothers had completed eighth grade or lessscored in the 29th percentile on the AFQTs. Thosewhose mothers had completed high school hadan average percentile score of 54. Those whosemothers were college graduates averaged 71.

Social class, according to Sue Berryman, "seemsto be a proxy for family characteristics that af-fect school achievement." These characteristics in-clude the family's press for achievement, lan-guage models in the home, academic guidanceprovided by the home," work habits and activi-ties of the family, and the "nature and quality oftoys, games, and hobbies available to the child."'°These characteristics correlate very highly withchildren's achievement scores. Later in life, socialclass correlates strongly with the level of resourcesa college student can bring to bear on his or hereducation. Financial problems in college are a fac-tor cited by many analysts in explaining the poorpersistence of minorities in the scientific and engi-neering educational system.

Despite the formidable socioeconomic problemsresponsible for the low participation rates inscience and engineering among blacks, Hispanics,and American Indians, there is considerable anec-dotal evidence that well-designed interventionprograms can assist these groups in obtaining ac-cess to science and engineering careers. The Amer-ican Association for the Advancement of Science,in an assessment of more than 100 such programs,found them to "have demonstrated that there areno inherent barriers to the successful participa-tion of women and minorities in science or math-ematics" if these groups are provided with "early,excellent, and sustained instruction in these aca-demic areas."5° These programs help stimulate aninterest in science and engineering among talentedstudents who might never have considered a ca-reer in one of those fields. They also provide sup-plementary instruction in science and mathematicsto help improve the academic preparedness ofthose students who choose to major in a quan-titative field. To date, programs for minoritieshave not been rigorously and systematicallyevaluated to determine the ingredients of successand failure. (Programs for women appear to have

"Berryman, op cit , p 15"Shirley M Malcom, Equity and Excellence Compatible Goals,

Publication 94-14 (N'ashington, DC American Association for theAdvancement of Science, December 1984), p yin

7

29

been more carefully studied.) Nor has any attemptbeen made on the national level to share the les-sons learned from successful programs with lessfortunate endeavors. NSF could take a leadershiprole in this area, but thus far it has not done so.In fact, the executive branch's response to themand,-:te it received from Congress in 1980 to de-velop a comprehensive program to promote theparticipation of women and minorities in scienceand engineering was to cut back or eliminate thefew programs in existence in that era.s'

Despite the problems cited above, women's par-ticipation in science and engineering has increaseddramatically in all fields and at all levels. Thereappear to be no inherent reasons why these in-creases should not continue. For blacks and His-panics the causes of low participation are sodeeply entwined with larger social and culturalfactors that the prospects for further improve-ments without dramatic societal intervention donot appear very bright. Already the rate of in-crease in participation among these groups hasslowed significantly since the dramatic improve-ments of the mid-1970s.

Finding 10

The increasing participation of foreign nationalsin U.S. Science and engineering education hasraised a number of unresolved policy issues.

The number of foreign nationals enrolled inAmerican institutions of higher education has in-creased by a factor of 10 in the past three dec-ades, from 34,000, or 1.4 percent of the studentpopulation in 1954, to 338,000, or 2.7 percent ofthe student population, in 1984." Foreign students

"The National Science Foundation Authorization and Science andTechnology Equal Opportunities Act of 1980 (Public Law 96-516,94 Stat. 3007) required the Director of NSF to prepare and transmitto Congress "a report proposing a comprehensive and continuingprogram at the Foundation to promote the full participation of mi-norities in science and technology" (Section 34b) The NSF responseto this requirement, "Proposals of the National Science Foundationto Promote the Full Participation of Minorities and Women inScience and Engineering," Dec. 15, 1981, typesrrrit provided byNational Science Foundation, contains the fPllowing quote

The Foundation's authorizing legiclaticn for fiscal year 1081 pro-vided for several programs that addressed the participation of minor-ities and women in science and technology These programs were re-duced by rescissions in fiscal year 1981 and were deleted from theFoundation's budget request for fiscal year 1982 as Fat of the Admin-istration's program for economic recovery

Two reports required of the President by Section 35 of Public Law96-516 were never transmitted to Congress"Open Doors: 1983/84 (New York Institute of International Edu-

cation, 1984), table 1.3, p. 6

tend to enroll proportionately more often in engi-neering and medical sciences than in humanitiesand social sciences. This has been attributed tothe relatively high cost to a developing countryof establishing proper training facilities fox theengineering and natural science disciplines. A pro-jected increase of 350,000 or more foreign collegestudents over the next decade could compensateto some degree for the expected decline in Amer-ican 18 to 24 year olds.

Foreign nationals have assumed a large shareof both enrollments and doctoral degrees in grad-uate departments of engineering, mathematics,and computer science (more than 40 percent ofenrollments and degrees in engineering; 40 per-cent of enrollments and 25 percent of degrees inmathematics). There is concern that becausescience and engineering education is capita! inten-sive, and many science and engineering graduatestudents are supported by Federal research as-sistantships, that the United States is subsidizingthe training of non-U.S. citizens. However, it ap-pears that many graduate engineering depart-ments would have had to curtail their researchand teaching activities substantially without for-eign student enrollments. Moreover, analystsLewis Solmon and Ruth Beddow have found, ina study of 1980-81 foreign degree recipients at thebachelor's, master's, and Ph.D. levels, that "for-eign students probably pay more than their edu-cation costs, and they might be paying substan-tially more."" Some also argue that there areforeign policy and good will benefits to U.S. edu-cation of foreign citizens. The U.S. Agency forInternational Development recently announcedthat it would increase its scholarship support toLatin American students in order to counter So-viet leadership in this area; this may be seen asevidence for this point of view.

There is disagreement between various groupswithin the United States as to whether individ-uals admitted to the United States as students whocomplete a graduate degree here should be al-lowed to remain here after completion of thedegree and perhaps a short period of further train-ing. Increasing proportions of those earning doc-

"Foreign Student Flows (New York Institute of International tit,-cation, 1985), pp 78-79

"Lewis C Solmon and Ruth Beddow, "Flows, Costs and Bene-fits of Foreign Students in the United States Do We Have a Prob-lem," Foreign Student Flows, op tit , pp 121-122

0 6

90

torate degrees (56 percent in 1983) are remainingin the country after graduation." Many employers,both in industry and in academic institutions, saythat they cannot fill certain research and facultypositions with U.S. citizens, and available dataverify this." Th.: Institute of Electrical and Elec-tronic Engineers, on the other hand, contends thatincreasing numbers of foreign engineers earningdegrees in this country are taking jobs at low payin order to remain here, thus driving down sala-ries and reducing opportunities for U.S. engi-neers." The limited statistical evidence that ex-ists, however, indicates that foreign engineers arenot paid less than comparably trained and experi-enced U.S. engineers."

Many foreign nationals choose to enter the aca-demic market after graduation, especially in engi-neering. A 1982 survey found that 18 percent of

"Participation of Foreign Citizens in U.S. Science and Engineer-ing (Washington, DC: National Science Foundation, January 1985),p. v.

"Betty Vetter, "U.S. Education and Utilization of Foreign Scien-tists and Engineers," contractor report prepared for the Office ofTechnology Assessment, p. 16.

"See statement of Jack Doyle, Institute of Electrical /ElectronicEngineers, in Scientific Manpower Commission, "The InternationalFlow of Scientific Talent: Data, Policies and Issues," May 7, 1985,

P. 29."Michael G. Finn, "Foreign National Scientists and Engineers in

the U.S. Labor Force, 1972-1982," draft report to the NationalScience Foundation, March 1985

all full-time engineering faculty members earnedtheir B.S. degrees from an institution outside theUnited States. Several authors have made refer-ence to a "growing problem" of foreign nationalson engineering faculties. Analysts contend thatthe drawbacks of an increasingly foreign bornfaculty are the limitations in their communicationsskills and difficulties they may have in adjustingtheir teaching and research styles to the expecta-tions of American culture."

Foieign nationals are often unable to work onindustry or government projects with nationalsecurity or economic competitiveness implica-tions. However, those students that obtain posi-tions in industry after graduation can and oftendo I".:come U.S. citizens arid thereby cease to be"foreign nationals."

Finally, some analysts contend that a lack ofdata prevents the Nation from making informedpolicy decisions in this area. However, there isa great deal of information about foreign studentsavailable from the Institute of International Edu-cation, and several new research projects spon-sored by NSF promise to fill many of the gaps.What is most lacking is a consensus over the def-inition of the "problem."

"The Internatonal Flow of Scientific and Technical Talent, op cit.

Chapter 2

Demographic Trends:Undergraduate and

Graduate Education

,1 lJ

Contents

PageThe Changing Composition of the Undergraduate Population 33Baccalaureates . 34Future Trends: Demographics .. 36Implications for Science and Engineering Personnel and Education Policy . . 38Graduation Education . 40Federal Support for Graduate Education in Science and Engineering ... 44The Future

. 50

List of TablesTable No Page2-1. Full-Time Equivalent Enrollments 362-2. Transition of 1980 and 1981 Science/Engineering Graduates From School to

Work 1982... 402-3. Percent of S/E Doctoral Recipients Who Were Foreign Nationals on

Temporary Visas, Selected Years . 442-4. Distribution of Primary Sources of Support for Full-Time Science and Engineering

Graduate Stude. '3 in Doctorate-Granting Departments, Selected Years . 452-5. Federal Predoctoral Fellowships and Traineeships by Field, Fiscal Years 1961-75 . 482-6. Full-Time Graduate Students in Doctorate-Granting Institutions Supported by the

Federal Government by Broad Field, Selected Years . . . 49

List of FiguresFigure No Page2-1. Bachelor's Degrees Conferred in Selected Fields. United States,

1965-66 ti 1980-81 342-2. Bachelo/s and First-Professional Degrees in Science and Engineering 352-3. Projected Full-Time Equivalent Students in Higher Education . . 362-4. Minority Enrollment As Perce-it of Public Elementary/Secondary School

Enrollment, by State . ... . . . 382-5. Projected Changes in Graduates, by State, 1981-2000 . . 392-6. Full-Time Graduate Enrollments in Science/Engineering at U S. Institutions,

by Field and by Year, 1960-83 . . . 422-7. Number of Doctorates Awarded by U.S. Institutions, by Field and by Year,

1958-83 . .. . .. . . . .. . 422-8. Doctorate Degrees in Science and Engineering by Field and by Year, 1958-83 432-9. Percent of S/E Ph D.s darned by Women .. 44

2-10. Federal Predoctoral Fellowships and Traineeships by Field, Fiscal Years 1961-75 472-11. Full-Time Graduate Students in Doctorate-Granting Institutions Supported by

the Federal Government by Broad Field, Selected Years . . 502-12. Doctorates Awarded . .. . . .. 512-13. Full-Time Graduate Enrollment by Area of Science/Engineering . . 512-14. Trends in Doctor's Degrees Conferred, by Sex of Recipient . . 52

Chapter 2

Demographic Trends:Undergraduate and Graduate Education

THE CHANGING COMPOSITION OF THEUNDERGRADUATF POPULATION

Significant changes in the size, composition,..nd career interests of the undergraduate popu-lation have characterized the past quarter centuryof American higher education. Demographic andother trends indicate that similar changes, al-though not as rapid or extreme, will also markthe next 20 to 30 years. Since the baccalaureateis the entry -level degree to 2 scientific or engineer-ing career, the composition of the undergraduatepopulation can be expected to have a significanteffect on the size and make-up of the science andengineering work force. This section will exa.-n-ine the changing demographic and career patternsof the undergraduate population and discuss theimplications of the trends for the size and qual-ity of the scientific and engineering work force.

The explosion in higher education that tookplace in the 1960s saw undergraduate enrollmentsincrease by 120 percent, from 3.3 million in 1960to 7.4 million in 1970.' This sizable change wascaused only in part by demographics. The num-ber of 18 to 24 year olds, who constituted about70 percent of the college population,2 also in-creased over the decade, but at a substantiallylower rate of about 54 percent.' Close to half ofthe increase in enrollment in the 1960s was causedby an increase in the rate at which high schoolgraduates chose to participate in higher education.In 1960 only 23.7 percent of the 18 to 24 year oldsthat graduated from high school were enrolled incollege: by 1970 that ratio had increased to 32.7percent, where it remains today.'

In the 1970s, undergraduate enrollments in-creased by 47 percent (from 7.4 million in 1970

i.:0 lical Abstract, 1982-1983 (Washington, DC. U S Departme Commerce, Bureau of the Census, 1984), p 159, corrected

. : .. :: unclassified pre-baccalaureates, and The Condition of Edu-.atio, ,.`fashington, DC National Center for Education Statistics,1985, :, 98

2Tne Condition of Education (Washington, DC National Cen-ter for Education Statistics, 1984), p 72

'Projections of the Population (Washington, DC U S Depart-ment of Commerce, Bureau of the Cens' s, 1984), p 7

to 10.9 million in 1983),5 but the population of18 to 24 year olds only increased 22 percent. Mostof the new college students were adults 25 yearsand older who doubled their participation inhigher education as a whole (including graduateschool) in that period. More than 70 peicent ofthe new over-25 group were adult women, whoseenrollment in higher education more than tripledbetween 1970 and 1983.6 The overall share ofwomen in higher education increased from 41.2percent in 1970 to 51.7 percent in 1983.7 Blacksincreased from 6.9 percent of the higher educa-tion population in 1970 to 10.6 percent in 1978,but then fell back to 9.6 percent in 1982. Hispanicsincreased from 2.0 to 4.4 percent between 1970and 1982; Asian Americans from 1.0 to 2.7 per-cent .s

In sum, from 1960 to 1983, the number of un-dergraduates increased by a factor of 3.:), from3.3 to 10.9 million. About 60 percent of thatchange was due to demographics: a 1.9-fold in-crease in the number of 18 to 24 year olds. Theremainder was caused by an increase in the rateat which high school graduates, especially adultwomen and minorities, chose to participate inhigher education. These enrollment trends tendedto favor public over private institutions and 2-year colleges over 4-year colleges. Enrollments inpublic institutions increased by a factor of 4.2(from 2.3 to 9.7 million) in the 23-year period;enrollments in 2-year colleges increased n-,ore thansevenfold (from 650,000 to 4.7 million)

'Statistical Abstract (Washington, DC U S Department ofCommerce, Bureau of the Census, 1985), p 149

'Condition of Education, op 'zit , 1985, 0 79°Ibid'Ibid , table 2 4, p 94"Statistical Abstract, op at , 1985, p 152°Statistical Abstract, 1982-83, op cit , p 159, corrected to in-

clude non-degree credit students, and Condition of Education, op.cit , 1985, p 79

'.33

34

BACCALAUREATES

The doubling in the enrollment of undergradu-ates in the 1960s was accompanied by a doublingin the number of bachelor's degrees awarded inthe course of the decade. In 1960, 395,000 stu-dents received B.A.s and B.S.s from the Nation'scolleges and universities; by 1970 that number hadincreased by 112 percent to 840,000.10

The 1970s saw a very different pattern. Al-though the number of undergraduates increasedby 47 percent between 1970 and 1983, the num-ber of baccalaureates only increased by 16 per-cent (from 840,00C., to 970,000)." This number ofawards below the bachelor's degree g ,nted u. oc-cupational curricula (business and corlmerce tech-nologies, data processing technologies, healthservices and paramedical technologies, 1. ci.ani-

"Statistical Abstract, 1979 (Washington, DC- U S Departmentof Commerce, Bureau of the Census, 1979) p 168, Condition ofEducation, op. cit., '985, p 90

"Condition of Education, op. at , 1985, p 90

cal, engineering, and electronic technologies) in-creased 160 percent (from 154,000 to 400,000)."Thus baccalaureates decreased from 84.5 to 70.4percent of the total undergraduate degrees awardedin the 1971-82 time frame.

T1' trend toward occupationally oriented degreesis nirrored by the changing mix of fields withinthe bialaureate in the 1970s As figure 2-1shows, the number of bachelor's degrees in job-related fields, especially business and manage-ment, increased dramatically in the 1970s, whilearts and sciences degrees declined substantially."According to the National Center for EducationStatistics (NCES): 14

"Sue E P ryman, 'The Adjustment of Youth and EducationalInstitutions Technologically-Generated Changes in Skill Require-ments," prepared for the National C.Anmission for Employment Pol-icy , May 1985, pp 30 and 31

"Digest of Education Statistics, 1983-1984 (Washington, DC Na-tional Center for Education Statistics, December 1983' p 119

"'bid , p 118

Figure 2.1.Bachelor's Degrees Conferred in Selected Fields: United States, 1965-66 to 1980-81

200,000

150,000

100,000

50,000

0

Arts r nd sciences Job-related fields

1965-66 1970-71 1975-76 1980-81 1965-66 1970-71 1975-76

SDURCE U S Department of

Year Yeardon. National Center for Education Statistics. Digest of Education Statistics 1983-84 p 119

43

1980-81

35

In 1965-66 degrees in English and literature, his-tory, mathematics. modern foreign languages,and physics constituted 20.4 percent of all bach-elor's degrees conferred . . . . by 1980-81 theyhad declined to 7.4 perceiit of the total . . . .

Ii. 1965-66 degrees in business and manage-ment, engineering, the health professions, publicaffairs and services, and computer and informa-tion sciences comprised 22..6 percent of the totalbachelcr's degrees conferred . . . . in 1980-81RI .w were] 41.8 percent . . . .

At least one occupationally oriented field hasnot participated in the general rise in recent years.Bachelor's degrees in education . . . peaked at194,000 in 1972-73 and subsequently declined by44.3 percent to 108,000 in 1980-81. There has beenan annual decrease in public school enrollmentsfor a decade, and this has adversely affect( 1 thedemand for new elementary and seconaary schoolteachers.

At first glance it appears that science and engi-neering fields have been immune from these strik-ing changes in undergra2uaLe career preferences.According to the National Science Foundation(NS)," science and engineering fields representedappkoximately 3A percent of the bachelor's andfirst professional degrees awarded every year from1952 to 1982. In no year have they exceeded 32percent or fallen short of 28 percent (Unfortu-nately, the NSF tables and charts include firstprofessional degrees, such as the M.D., D.D.S,D.V.M, and J.D., with .-;;Achelor's degrees, co thenumbers are not strictly comparable tt., those pre-sented above.) The total number of science andengineering bachelor's degre ^ and first profes-sional degrees has retrained within the 280,000to 305,000 range sinco 1472.

This apparent L.instancy, however, masks somestriking changes ta-ing place within and betweenfields (see figure 2-2). From 1960 to 1974 the per-centage of total B.A.c anA first professionaldegrees awarded in the pt,,--ical sciences and engi-neering fell from 4.1 and 9." " ercent, respectively,

"Science and Engineering Personnel. A National Overview, NSF85-302 (Washington, DC: hi-ional Science Foundation, 1985), p154.

figure 2-2.Bacnelor's and First-ProfessionalDegrees in Science and Engineering

320

300

280

280

240

220

200

180

150

140

120

100

80

80

40

200

1985 1970 1975 1980 1982

YearSOURCE National Science Foundation, Science and Technology Resources,

NSF 85-305, 1985, p 73

to 2.1 and 4.3 percent. That decline was accom-panied by an increase in the percentage of degreesawarded in the social sciences from 8.0 to 14.4percent. Between 1974 and 1982 the trend reverseditself, and engineering degrees increased from 4.3to 6.5 percent of the total, while social sciencedeg. ees fell from 14.4 to 10.9 percent."'

Science and engineering baccalaureates, there-fore, appear to follow the same trends as total bac-calaureates; increasing with total college enroll-ments through the 1960s and then going up onlyvery gradually in the 1970s a.-id early 1980s. In-dividual fields, however, follow changes in un-dergraduate interests, such as the shift from thenatural sciences to the social sciences in the 1960sand from academic subjects to job-oriented sub-jects in the 1970s. It is a combinatiod of demo-graphic forces and career and educational choices(strongly influenced by market considerations)that determines the number of bachelor's degreesthat will be produced in a science or engineeringfiele in a given year.

i6lbid

i

36

FUTURE TRENDS:

According to demographer Harold Hodgkin-son, three major demographic trends will glee'colleges and universities over the next quarter cen-tury. The first is the already cited decline in thepopulation of 18 to 24 year olds:17

Higher education can look forward to a gen-eral decrease in the size of high school graduat-ing classes for 16 years, followed by increasednumbers of high school graduates beginning in1998 . . . .

Most recent Census Bureau projections (MiddlSeries) indicate that the number of 18 to 24 yearolds will decline from 30.4 million in 1982 to 23.3million in 1996, a 23.4-percent decrease." Thisis the group that participates at the highest ratein higher educaticn. As can be seen from table2-1, for every 100 members of the 18- to 21-year-cld population there are 27 full-time equivalent(FTE) students in higher education. (The numberof "full-time equivalent students" is equal to thenumber of full-time students plus one-third thenumber of part-time students.) The mt.() for 22to 24 year olds is 11.5, and the ratios for all otherage groups are much lower."

Using the ratios in table 2-1 and the Census Bu-reau Middle Series projections of the populationby age group,2° it is relatively straightforward to

"Harold Hodgkinson, "College Students in the 1990's A Demo-graphic Portrait," The Education Digest, November 1983, pp 28-31

"Projections of the Population, op cit , pp 38 and 65"Carol "erring and Allen R. Sanderson, Doctorate Employment

Supply ar Demand Considerations, 1981-2000 (Princeton, NJPrinceton University, 1981), p 8

"Projections of the Population, op at , pp 30-81

Table 2-1.F ull-Tima Equivalent Enrollments

As a percentage of the total populationAge group 19680 1973a 1978a 1980-2000b17 .... . . 71% 6 7% 58% 60%18-21... ... 28.0 28 0 26.0 27 022-24 9 7 11 2 11 4 11 525-2£ . .. 4 0 5.0 5 0 5 030-34. ... 1.9 2 5 3 0 3 035-50 0.4 0 6 0 8 0 8,Actual°ssumed, based on current trends

SOURCE Carol P Herring and Allen R Sanderson Doctoral Employment Supply end Demand Considerations. 1981 2000, Princeton University April1981, p 8

lr

DEMOGRAPHICS

construct a table of potential FTE enrollments inhigher education for the rest of the century (andeven to the year 2020, if one is willing to acceptpopulation projections based on births that havenot occurred yet). Figure 2-3 shows a continueddecline in FTE enrollments through 1996, with theminimum in that year at 85 percent of the peakin 1981. Enrollments then increase through thefirst decade of the 21st century, with a new peakoccurring in 2010 about 15 percent above the 1996minimum."

These projections are based on a model devel-oped by Carol Herring and Allen Sanderson thatassumes existing patterns of enrollment in highereducation among different age groups woo con-tinue for the next quarter century. Other as.,ump-tions could produce strikingly different results.For example, the number of 25 to 34 year oldsin under raduate education increased by nearly70 percen. ;n the decade 1972-82, as compare d toan increase of only 23 percent in the number of18 to 24 year olds. If this older group, which willremain relatively constant in population over thenext decade, continues to increase its rate of par-ticipation in undergraduate education, it couldpartially offset the decline in 18 to 24 year old en-

"FTEs through 2000 calculated by Herring and Sanderson, opcit , FTEs, 2000-20 calculated by Eugene Frankel.

Figure 2-3.Projected Full-Time EquivalentStudents in Higher Education

10,000

8,000

6,000

4,000

2,000

0

1980 1984 1988 1992 1998 2000 2004 2008 2012 2018 2022

iub

YearSOURCE Carol P Fleming and 4tten R Sanderson, Doctoral Employmen

Supply and Demand Considerations, 1981-2000 Princeton Unites,ty April 1981 p 8 and Office of Technology Assessment

37

rollments.22 Specifically, if one assumes that thepercent enrollments in higher education of 30 to34 year olds and 35 to 59 year olds will continueto increase at the rates they increased between1968 and 1978 (0.1 and 0.04 percent per year),it would generate an additional 1 million full-timeequivalents in 1996. The decline of 15 percent inFTE5 between 1981 and 1996 would become a muchless problematic decrease of only 2.4 percent.

;-lerrng and Sanderson argue against thesemore optimistic assumptions:"

There are those who . . . argue that the growthin "non-traditional" enrollments will more thancompensate for the loss of the traditional popu-lation base in higher edw-ation. We disagree. . . . The recent increase (in non-traditional en-rollments] has been due in large measure towomen going back to school, and at the presenttime they make up about half of the enrolled stu-dents. Therefore, the impact of this catching upphenomenonin which women were making upfor missed educational opportunitiesis expectedto level off . . . . Part of the increase in older stu-dents can also be traced to the effects of the Viet-nam War and the GI Bill. These effects, too, haveundoubtedly run their course, and most veterarcwho are going to return to school on the GI Billhave done so by now.

The NCI2.S projection of enrollments in highereducation for 1993 is consistent with Herring andSanderson's results. It projects a drop in FTE en-rollment; of 12.3 percent between 1983 and1993." Herring and Sanderson also project a dropof 12.3 percent. It is worth noting that in theNCES projections the fraction of enrollments thatare part time increases from 42.9 percent in 1983to 48.3 percent in 1993.

The second major trend noted by Hodgkinsonis the increasing proportion of the college age pop-ulation that will be made up of racial or ethnicminorities:"

The major decline in births after the baby boomwas almost completely a Caucasian phenomenon.Birthrates for minorities stayed even clufng thoseyears, resulting in an increased percentage of

"Digest of Educational Statistics, 1983-1984, op cit , p 98"Herring and Sanderson, op cit , p 6"The Condition of Education, op cit , 1985, pp 96 and 100"Hodgkinson, op cit., p 29

births coming from minorities . . . . The conclu-sion for higher education is inescapable: Amer-ican public schools are now very heavily enrolledwith minority students, large numbers of whomwill be college eligible . . . . Out of sheer self-interest, it behooves the higher education com-munity to do everything to make sure that thelargest possible number of minority students dowell in public school . . . . By 1990, minorities ofall ages will constitute 20 to 25 rercent of our total population, while their percentage amongyouth cohorts will be over 30 percent. In somestates, minorities will be over 45 percent of thestate birth cohorts.

Public school enrollments in 1980 presage thesetrends. Of the 46 million students enrolled in theNation's public elementary and secondary schoolsin 1980, 26.7 percent were minority. Of the largestminority groups, blacks represented 16.1 percentof enrollments, Hispanics, 8.0 percent." Se-fenStates and the District of Columbia had minor-ity student populations in excess of 35 percent;11 others exceeded 25 perceni (see figure 2-4).27In addition, 23 of the 25 largest city school sys-tems in the Nation had a majority minority pop-ulation in 1980.2°

The report of a 1983 forum on "The Demo-graphics of Changing Ethnic Populations andTheir Implications for Elementary-Secondary andPost-secondary Educational Policy" concludesthat:"

Based on the demographic data . . . it seemsclear that much greater attention will have tc, bepaid to the needs of minority young people, andto the development of programs that are more re-sponsive to their backgrounds and interests, forfacilities and equipment to susta'n these pro-grams, and for teachers specifically trained toteach particular populations, . . . More researchis needed into how young people of differentbackgrounds learn, and existing research shouldbe mined and adapted for practical application toloca: conditions.

The Condition of Education, op cit , 1985, p 18."Ibid , pp 18 and 19"Ian McNett, "Demograph:c Imperatives- Implications for Educa-

tional Policy," Report of the June 8, 1983, Forum on The Demo-graphics of Changing Ethnic Populations and Their Implications forElementary-Secondary and Postsecondary Educational Policy," p.10

"Ibid , p 20

38

Figure 24.Minority Enrollment As Percert of Public Elementary/Secondary School Enrollment, by State

F.L"Fil IP

SOURCE Natioial Cantar for Education Statistics, Tn. Condition of Education, 1984 Edition p 19

Percent minority in 1980

ME s'

NHMA

T RI

NJ

DE

0 Less than 25.0 percent minority

0 25.0 to 34 9 percent mincity

' III 350 or more percent minority

\)

A third trend noted by Hodgkinson and othersis the striking variation in population growth fromregion to region. As can be seen from figure 2-5,30 the number of high school graduates is ex-pected to decrease between 1981 and 2000 in theNorth, East, South, and Central sections of thecountry, while increasing in the West, Far West,and Southwest. Moreover, some of the increases(in Wyoming, Nevada, Utah, Texas, and Alaska)will be as large as 48 to 76 percent, while someof the decreases (in Michigan, Massachusetts,New Hampshire, Rhode Island, Delaware, andDistrict of Columbia) will be as great as 30 to 51percent .31

"Harold L. Hodgkinson, "Demographics and the Economy Un-derstanding a Changing Marketplace," The Admissions Strategist,January 1985, pp. 1-6.

"Hodgkinson, op cit., 1983, p 29

Thus [according to Hodgkinson] we are facedfor the first time with a "two nation" perspectiveon educational policytrying to get more educa-tonal facilities and services for youth in the Sunbelt, and continuing to cut back en these same

'rvices in the Frost Belt. It is hard to imagine howa single federal policy on educational assistancecan be equitable in both "nations." Further, sincewe can expect that institutions of higher educa-tion in the Sun Belt will begin to show gains inundergraduate student populatiors, while theFrost Belt institutions can look forward to at leasta decade of decline in enrollments, it will be verydifficult to present the "needs" of higher educa-tion to the U.S Congress by 1990.

IMPLICATIONS FOR SCIENCE AND ENGINEERING PERSONNELAND EDUCATION POLICY

The implications of these trends for science andengineering education appear to be the following:As the number of college graduates available for

science and engineering careers decreases, it willbecome increasingly important to ensure that allpotential resources are utilized t:-1 the fullest ex-

47

39

Figure 2.5. Projected Changes In Graduates, by State, 1981-2000

Down22%

Down 5%aDown 191/8abDown 35%abDown 310 /tabDown 31%Down 29%aDown 20%Down 38%Down 51%a

III Gain0 Loss of 1% to 250/.

0 Loss of 280/s or more

"'Projections Include graduates of public e schools only All other State protections include public and nonpublic high school graduatesbProjections are for tellegg

SOURCES. western Interstate Commission for Higher Education, and Harold L Hodkinson, "Demographics and the Economy UnderstandingChanging Marketplace," The Admission Strategist, January 1985, p 3

tent possible, and that no qualified candidates forscience and engineering education are discouragedby problems related to gender, race, or ethnicbackground. Hence, policies to promote equal-ity of opportunity for women and minorities willtake on increasing relevance to science and engi-neering "manpower" policy.

Blacks and Hispanics may be of special concern.They tend to participate in higher education athalf the rate of their white count:Tarts, and tendto select quarfi;tative fie:ds as majors once in col-lege or graduate school at one-half (for blacks)to three-fourths (for Hispanics) the rate of thewhite student population." As the fraction of thecollege age population increases, programs to pro-mote equalit of access to science and engineer-

"Sue E. Berryman, Who Will Do Science (New York. The Rocke-feller Foundatio", 1983), pp. 18-21.

ing careers for minority groups may have to beexpanded in order to prevent a decline in the num-ber of science and engineering degrees awarded.These programs will be discussed in greater de-tail in chapter 5.

It is not possible to judge unequivocally the ef-fects of a drop of 15 percent in the number ofscience and engineering baccalaureates on the Na-tion's scientific and technical efforts. NSF data dis-played in table 2-2 show that only 43 percent ofthe science and engineering baccalaureates in 1980and 1981 were employed in a science or engineer-ing field in 1982.33 Of the remainiag science andeng ering baccalaureates, 28 percent foundwork outside scir ice and engineering altogether,21 percent went on to grauuate school in science

"National Patterns of Science and Technology, 1984, NSF 84-311 (Washington, DC National Science Foundation, 198t), p. 24.

40

Table 2.2. Transition of 1980 and 1981 SciencelEngineerIng (SIE) Graduates From School to Work in 1982 (percent)

Degree level/field TotalEmployed Full-time

graduatestudents

Not employedSeeking

employmentOutside

labor forceTotal In S/E Outside S/EBachelor's dogrm:

All S/E 100 71 43 28 21 5 4Physical sciencesa 100 53 41 12 40 4 3Mathematicsuomptster specialties

100100

8(,94

5785

239

133

42

41Engineering 100 89 78 10 7 3 2Ufe sciences 100 58 34 24 33 5 5Social sciences° 100 68 21 46 22 6 5

Wastes dorm:All S/E 100 74 55 19 21 3 2Physical sciencesa 100 65 47 18 29 3 3Mathematics 100 77 56 21 17 3 3Computer specialties 100 88 76 12 10 1 1Engineering 100 82 72 11 14 2 2Ufe sciences. 100 65 47 17 31 2 3Social sciences° 100 63 34 34 24 6 3

'includes envirormental sciencesI/Includes psychology

SOURCE. National Science Foundation, National Patterns of Science and Technology Resources, NSr 84-311 , p 24

and engineering and other fields and another 5percent wer.: still seeking employment. Thus, itappears th tt the Nation is currently producingmore science and engineering baccalaureates thanare absolutely required by the science and engi-neering work force.

This finding clearly does not hold true for engi-neering and computer science, which ,bowed em-ployment rates in science and engineering of about80 percent. However, engineering and computersciences illustrate another important feature of thescience and engineering marketplace: the extremeresponsiveness of undergraduates to market sig-nals and career opportunities. Undergraduate en-rollments in both computer science and engineer-ing have increased dramatically in the last 8 yearsin response to excellent job opportunities for bac-calaureates in those fields, as discussed in chap-ter 1. If other scientific fields were to require

greater rr,mbers of technically trained baccalaure-ates, market signals alone would attract an in-creasing supply of undergraduates to them. AsSue Berryman has found:"

Overall, the data on the levels and fields ofcompleted degrees suggest that youth respond tooversupplies by earning fewer degrees in oversup-plied fields. If they enter an oversupplied field,they increase the amount of education they ob-tainpresumably to increase their competitive-ness in a loose labor ma,ket. They respond toshortages or more liberal employment opportu-nities by increasing their educational investmentsin these fields at the lower degree levels and re-ducing them at the higher degree levelspresum-ably because they are in a seller's market.

"Sue E Berryman, 'The Adjustment of Youth and EducationalInstitutions to Technologically-Generated Changes in Skill Require-ments," op cit., p v

GRADUATE EDUCATIONGraduate school is the principal training ground

for a research career. Here the student gains theadvanced knowledge and techniques that are thetools of the research trade, and forms the profes-sional contacts that shape a research career. Agreat deal of detailed information about science

and engineering graduate education is availablefrom the National Science Foundation, the Na-tional Aradem} of Sciences, the Institute of Medi-cine, and the science and engineering professionalsocieties. Science and engineering graduate stu-dents are classified by field, degree, source of sup-

4 ,3

port, type of institution, sex, racial and ethnicgroup, and citizenship. Ph.D. scientists are prob-ably the most thoroughly surveyed and well-documented citizens in the United States. This sec-tion summarizes the principal trends in graduateeducation, focusing on the doctoral degree as tit?entry level to a research career.

Graduate enrollments have followed a patternsimilar to that of unrergraduates in the past qaar-ter century, dramatically increasing in the 1960sand then rising more modestly in the 1970s andearly 1980s. However, the relationship to largerdemographic trends is far more tenuous than inthe undergraduate case. From 1960 to 1970 grad-uate enrollments nearly tripled, rising from360,000 to over a million, while at the same time,the population of 22 to 34 year olds increased byonly 18 percent.35 From 1970 to 1982, by contrast,the number of graduate students :ncreased onlymodestly, by 30 percent, while the overall popu-lation of 22 to 34 year olds rose more than 55 per-cent." (See below.) Thus there appears to be nodirect relationship between graduate enrollmentsand the graduate school age population.

Science and engineering have followed overallgraduate enrollment trends quite closely. From1960 to 1970 the number of full-time graduatescience and engineering students enrolled in U.S.institutions increased 150 percent, from about78,000 to 188,000." By 1982 that number had

"Statistical Abstract, 1982-83. op. cit. , pp 25 and 150"Condition of Education, op. at , 1985, p 79, Protections of the

Population, op. cit , p. 38"Susan Fallows. "Federal Support for Graduate Education in the

Sciences and Engineering," background paper prepared for the AdHoc Committee on Government-University Relationships in Sup-port of Science, National Academy of Sciences, 1983, p. 10.

41

reached 264,000, a further increase of 40 percent."(An additional 150,000 science and engineeringgraduate students were enrolled part time, bring-ing the total to 414,000.") Enrollment growth wasgreatest in the social sciences and the life sciencesover the full 23-year period, as can be seen fromfigure 2-6, and weakest in engineering, physicalscience, and especially mathematics and computerscience."

The number of doctorate degrees awarded an-nually in all fields tripled in the decade of the1960s, reaching a peak ii 1973 at 33,755. It hassubsequently declined to a level of 31,000 peryear, where it has remained since 1978 (see aboveand figure 2-7)." The number of science and engi-neering Ph.D.s grew slightly less rapidly, by a fac-tor of 2.85 between 1960 and 1970. It reached apeak at 19,009 in 1972 and has subsequently de-clined to a nearly constant level of 17,500 to18,000 per year in the 1980s.42 The share of totaldoctorates awarded in science or engineering de-creased from 65 percent in the early 196Gs to 55percent by the late 1970s."

"Academic Science, Engineering Graduate Enrollment and Sup-per Fall 1982, NSF 85-300 (Washington, DC National ScienceFoundation, 1925), p 20

"Ibid"Fallows, op u. , p 12"Ibid , pp 31 and 32, Science and Engineering Personnel A Na-

tional Overview, op cit , p 156, gives slightly different numbersfor total Ph Els 9,733 in 1960, 29,498 in 1970

"Fa lows, op cit , p 31 , Science and Engineering Personnel ANational Overview, op. cit., again gives d=fferent numbers, 6,263science and engineering Ph D s in 1960 and 17.743 in 1970. Thisis due to a difference in classification of life science and social science

fields between the National Science Foundation and the NationalAcademy of Sciences

"Science and Engineering Personnel A National Overview, op.cit.. p. 156

4

,1

42

Figure 2-6.FullTimi, Graduate Enrollments inSciencelEngineering at U.S. Institutions, by Field

and by Year, 1960-83

eo

so

2

4°

20

01950 1965 1970 1975 1980 1985

YearSOURCE Susan Fallows, Federal Suppor for Graduate Education 'he Sciences

and Engineering (Washington, DC National Academy of Sciences19831, p 12

The number of Ph.D.s awarded annually grewat roughly the same rate in all sciencr. and engi-neering fields during the 1960s, as can be seenfrom figure 2-8. in the early 1970s, however, thepaths of the different disciplines began to diverge.The number of doctorates awarded annually inengineering, and the mathematical and physicalsciences decreased by 30 percent between 1971 and1978, dropping from 4,500 to 3,200 per year inthe physical sciences, from 3,500 to 2,400 in engi-neering and from 1,300 to 960 in mathematics andcomputer science. Meanwhile, the number of lifescience Ph.D.s remained constant at about 4,400per year, while the number of social sciencePh.D.s increased from 5,000 to 6,000 per year .44Since 1978 the number of doctorates awarded each

"Fallows, op cit , p 33

Figure 2-7.Number of Doctorates Awarded byU.S. Institutions, by Field and by Year, 1958-8"

30

§20

x

a.'41

a.

10

6

01958 60 82 64 66 68 70 72 74 76 78 80 82 83

YearSOURCES Susan Fallows Federal Support for Graduate Education in the Sci-

ences and Engineering ,Washington DC National Academy of Sci-ences 19831 p 12 and National Research Council, Doctoral Recto'ents From United States Universities Summary Report, 1983, p 5

year in the various science and engineering fieldshas remained relatively constant, with modest in-creases taking place in engineering and the lifesciences.

The changes in graduate enrollments and an-nual doctorates awarded in the different scienceand engineering fields are related to broaderchanges in the composition of the graduate schoolpopulation since 1970. Women and foreign na-tionals have dramatically increased their sharesof both enrollments and doctorates over the past15 years, with the effects varying from field tofield. According to educational consultant RobertSnyder:"

It is important to note that graduate science andengineering enrollments would have remained

"R G Snyder, 'Some Indicators of the Condition of GraduateEducation in Science and Engireering, paper presented at TheBrookings Institution, Nov 15, 1984, typescript, pp 4-5

43

Figure 24Doctorate Degrees in Science andEngineering by Field and by Year, 1958-83

7.000

6.000

5.000

08 3.000

2.000

1,000

1960 1964 1968 1972 1976 1980 1985

SOURCES Susan Fellows, Federal Support for Graduate Education in the Sciences and Engineering (Washington, DC National Academy of Sciences, 1983), p 12, National Research Council, Doctoral RecipientsFrom United States Universities, Summary Report, 1983, p 5, and National Science Foundation, Science and Technology Resources NSF85-305, January 1985, pp 83-85

flat, were it not for the marked increases in fe-male and foreign enrollments. With respect togender difference , female full-time enrollment indoctoral institutions rose 53 percent in the 1975-1982 time period, compared to level enrollmentfor men. While sharp percentage increases oc-curred in engineering, environmental and com-puter sciences, these increases were built upon ameager base. Therefore, as of 1982, women com-prised only 11 percent of full-time engineering stu-dents, 20 percent in physical/en ironmentalsciences and 25 percent in mathematical/com-puter sciences. The greatest impact of female en-rollment increases ha., been in the biological

sciences (40 percent women) and psychology/so-cial sciences (46 percent women). In fact, maltfull-time enrollment in the 1975-82 period declined14 percent in the biological sciences and 17 per-cent in psychology/social sciences compared tofemale ilizreases of 38 and 35 percent, respec-tively, in the same period. Thus female enroll-ments were wholly responsible for positive enroll-ment trends in these two broad fields.

In [the "EMP" sciences], enrollment increaseshave been due significantly to surging enrollmentsof foreign students. 1 engineering, foreign full-time enrollments rose at an average annual rateof 8.4 percent f..om 1975 to 1982, reaching almost21,000 students or 43 percent of total full-time sttedents in this broad field. This compares to inannual increase among U.S. citizens of only 1.6percent. An even larger discrepancy was in math-ematics/computer sciences. Here, foreign enroll-ments increased at a rate of 12.7 percent per year,to a 36 percent share of full-time enrollments,compared to z 0.5 percent per year increase forU.S. citizens. One encouraging note is that in1981-82, the most recent year for which data areavailable, U.S. enrollments in both these broadfields rose at much higher rates than previously.

With respect to minority participation in grad-uate science and engineering education, the pic-ture in recent years is not encouraging. While ex-tended trend data are not available, recent NSFdata show very low rates of participation. For allfields, blacks comprise only 2.5 percent andHispanics 2.1 percent of full-time science and engi-neering enrollments in 1982.

Data on doctoral degree production confirmthese trends. Figure 2-9 shows the dramatic in-creases in percentage of Ph.D.s awarded towomen in all fields between 1965 and 1983.'6 Ta-ble 2-3 shows the dramatic increase in the per-centage of science and engineering doctoral recip-ients who were foreign nationals on temporaryvisas between 1970 and 1983 in physics/astron-omy, chemistry, mathematics, and engineering.The first two fields increased by more than a fac-tor of 2; the last two by a factor of 3.'7

"Opportunities in Science and Engineering (Washington, DC:Scientific Manpower Commission, 1984), p 8

"Professional Women and Minorities (Washington, DC Scien-tific Manpower Commission, August 1984), p. 32,

44

Figure 2-9.-Percent of S/E Ph.D.s Earnedby Women

40

30

20

O

10

01985 1970 1975 1980

Year

SOURCE Scientific Manpower Commission, Opportunities in Science endEngineering, Washington, DC, 1984, p 8

Table 2.3.- Percent of S/E Doctoral Recipients Who Were Foreign Nationals on Temporary Visas, Selected Years

Field 1962 1966 1970 1974 1980 1983Physicslastronomy 13 7% 12.2% 11.3% 17.2% 19.2% 24.8%Chemistry 9.7 11.1 7.9 10.2 15.4 18.1Earth/environmental science 9.2 12.9 13.9 16.1 12.7 18.8Mathematics 16.8 12.6 10.9 18.5 13.7 29.8Economics 17.2 20 1 17.2 22.0 25.0 25.9Political science 19.1 13.0 9.3 10.8 11.9 13.9Engineering 17.9 16.7 13.7 22.4 34.3 42.1Biological science 14.4 14.3 10.5 10.5 7.8 8.8Medical science 14 6 18.6 16.6 9.4 11.3 13.7Agricultural science 23.8 30 4 25 3 33.2 32.6 31.2Total S/E 15.2% 15.3% 12.5% 16.7% 18.8% 15,5%SOURCE: Scientific Manpower Commission, Professional Women and Minorities,Washington, DC, 1984, p 32

FEDERAL SUPPORT FOR GRADUATE EDUCATIONIN SCIENCE AND ENGINEERING

The pattern of graduate enrollment and Ph.D.production in sciel. :e and engineering in the 1960-75 era is also related to the changing nature andlevel of Federal support for graduate educationin those fields. Since World War II, the FederalGovernment has utilized a variety of mechanismsto support graduate students and further the goalof enhancing the Nation's resources of highlytrained scientists and engineers. Fellowships andtrainaeships are the direct forms of support forgradual:Attu:lents that have been most frequentlyused to target areas of perceived personnel need.Research grants and inst:aitional support havealso provided essential, though indir..ct, meansof supporting graduate education.

Fellowships are tuition/stipend mechanismsthat are awarded directly to students through na-tional competition. Award is made principally onthe basis of merit, the intent being to attract thebest students available into a particular field orproblem area. Fellowships are portable in thesense that the awardee is free to enroll in any qual-ified institution to which he/she can gain ad-mittance.

Federal traineeship programs (sometimes called"training grants"), in contrast to fellowships, areblocks of student tuition/stipend awards that aremade to departments or institutions on a multi-year basis. Traineeships focus Jr1 the ;raining pro-

45

gram, not the student. Departments typically ap-ply for training grants in national competition,and awards are made by peer review so that highquality standards are met. Because of the multi-year block nature of training grants, they are wellsuited to target personnel at social problem solv-ing areas as well as broad fields of science. TheNational Institutes of Health (NIH) traininggrants, historically the largest of these programs,have been used for both purposes.

Federal research grants are an important vehi-cle for training and support of graduate studentsthrough the research assistantship (RA). Since re-search training is generally acquired through anapprenticeship with a faculty member, the RAaids in this mentoring process, while providinga source of financial support for the student. How-ever, since the primary purpose of the RA is toassist in the production of research, the numberof RAs awarded is often not explicitly related tonational personnel and training requirements.

A final source of Federal support for graduateeducation is the institutional grar , such as thegeneral science support grant, facilities and equip-ment support, and the science development grant.Most of these forms of assistance do not supportgradua' e students directly, but undergird the re-search and education environment.

Prior to 1958, the primary source of Federalassistance to graduate education in science andengineering was the research assistantship. Ac-

cording to education consultant Robert Snyder,, "a survey of graduate student support in 1954 byNSF indicated that 10 percent of all graduate stu-dents received Federal stipend support (6,751), ofwhich 86 r?rcent came in the form of research as-sistantships." By 1969, the situation had changeddramatically. 36 percent of all graduate scienceand engin.... students (51,620) were supportedby the Government, and 56 percent ofthose supporied held either a fellowship or atrainceship. Table 2-4 illustrates those changes.

The impetus for this change was the launchingof Sputnik by the Soviet Union in October 1957.The following year Congress passed the NationalDefense Education Act (NDEA, Public Law 85-864) which "marked the inception of large-scaleFederal graduate education support."" The TitleIV fellowship program of NDEA grew to supportmore than 15,000 graduate students at a budgetof $86 million in 1968, with support being pro-vided in many academic fields, including theheretofore excluded humanities." In addition,NDEA also "provided subsidies to educational in-stitutions to create low interest loans for needystudents in all disciplines . . . . [which] would be

"Robert G Snyder, "Federal Support rf Graduate Education inthe Natural Sciences An Inquiry Into the Social Impact of PublicPolicy," doctoral dissertation, Syracuse .Inivers,-i, June 1981, pp.103 and 110, appendix tables 6 and 9 pp 264 and 267

"Ibid , p 108"Ibid

Table 2.4. Distribution of Primary Sources ^I Support for Full-Time Science and EngineeringGraduate Students in DoctorateGrarting Departments, Selected Years

Source of support 1954 1966 1969 1974 1979 1983

Number of graduate students:All sources ... 67,136 118,273 141,199 195,915 223,409 243,646Federal Government 6,751 44,612 51,620 48,016 52,871 47 402Institution /State 16,958 42,343 50,471 75,411 82,813 96,188Self . 39,000 19,571 26,307 56,085 37,686 75,641Other 4,427 11,747 12,801 16,403 29,039 24,415

Percent distribution:All sources 100 0 100 0 100 0 100 0 100 0 100 0Federal Government 10.1 37 7 36 6 24 5 23 7 195Institution /State 25 3 35 8 35 7 38 5 3/ 1 39.5Self 58.1 166 186 28 6 30 3 31.0Other 66 99 91 84 90 10.0NOTE Except for 1979 and 1983 absolute numbers in this table should be used with caution for trend analystr, because they are derived from somewhat Incompatible

sources 1954 data include part time students thereby overrepresenting self support

SOURCES National Science Foundation Graduate Student Enrollment and Support in American Universities and Colleges, 1951, National Science Foundatio Greduate Student Sr.pport and Manpower Resources In Graduate Science Education fall 1966 anu fall 1969, National Science Foundation, Academic SolanaGraduate Enrollment and Support Fall, 1978, and National Scie -e Foundation Academic Sciance,Engineering Graduate Enrollment and Support, Fall 1983

forgiven for students who later went into teach-ing careers."S1

In the years following the passac- of NDEA,the President's Science Advisory Committee is-suet a series of reports calling for increased sup-port for science and engineering education. ThesereportsEducation for the Age of Science (1959);Scientific Progress, the r iniversities and the Fed-eral Government (the "Seaborg report," 1960),and Meeting Manpower , in Science andTechnology (the "Gillilan report," 1962)"articulated the national need for greater numbersof scientists and engineers . . and for strongerFederal support for the training of manpower forbasic research and for university level teaching ""They were followed by a set of significant actions:

1. The NSF fellowship program, begun in 1952,was expanded from about KJ to 2,500 pryear.

2. An NSF traineeship program was begui, in1962, and awarded 5,000 traineeships a yearto 200 institutions by the late 1960s.

3. NIH predoctoral support programs expandedfrom 55 traineeships in 1958 to 7,696 trainee-ships and 1,527 fellowships in 1970.

4. The National Aeronautics and Space !Amin-istration (NASA) instituted a traineeshipplograrn in 1962 for engine, mathemati-cians and physical scientists.

5. The Atomic Energy Commission (AEC) ex-panded its fellowship program and ins'...uteda new traineeship program (1965) in nuclearscience and engineering.

All told, the number of science and engineer-ing graduate students supported by Federal pre-doctoral fellowships end traineeships increasedfrom 5,000 in 1961 to 34,100 in the peak year of1969.53 (See figure 2-10.) At the same time Fed-eral research and development (R&D/ support touniversities expanded fourfold, from $405 millionin 1960 to '4.595 billion in current dollars in1969.5' As a result, the number of federally sup-ported research assistantships increased to morethan 19,600 .n 1969.55

"Fallows, op. cit , r 5"Ibid."Snyder. op. cit 1981, p 109 and appendix table 7, p 265"Fallows, np. cit , p 404"Snyder, of cit., 1981, p 267 The number 19,646 given in ap-

pendix table 9 of Snyder's dissertation represents those full-time grad-

In addition to direct stipends for graduate studyand research, two new programs in the mid-1960sbroadened the scope of Federal support for grad-uate education. These were the Guaranteed Stu-dent Loan and College Work Study programs au-thorized by the Higher Education Act of 1965(Public Law 89-379). The work study programenabled gradua. departments and researchers

to hire kw ir.:ome students at a small fractionof the costs of 0- students' wages through Fed-i -al wage subsidies." The loan program wasdirected at low-income graduate and undergradu-ate students in all fields "as part of a larger shiftin Federal policy towards aiding disadvantagedsocioeconomic groups."" By 1969 more than70,000 students were receiving guaranteed studentloans; 17,000 were enrolled in college work-studypre' and 27,403 held NDEA loans. An addi-ti,nal 91,464 graduate and professional schoolstudents (the numbers are not available for grad-uate students alone) were supported by veteransbenefits." It should be noted, however, that theaverage level of support per student from theseprograms was less than $1,000 per year, as com-pared to more than $5,500 per year, tuition plusstipend, from the NDEA Title IV fellowshipprogram.

Thus, between 1960 and 1969 Fedel al supportfor graduate education in science and et.eineer-ing expanded dramatically, as part of in overallFederal emphasis on post-secondary education.Other forces were at work as we': during thisperiod. Increasing State support f it higher edu-cation, and especially the expansion (.. many Stateuniversity systems in this era, created at, increas-ing demand for new faculty to instruct the grow-ing numbers of undergraduates. And increasingundergraduate enrollments created significant ad-ditional demands for teaching assistants, therebyproviding much additional giadua. :,tudent sup-port. By 1969, over 50,000 full-time science andengineering graduate students were receiving theirprimary source of support from institutional andState s ...-ces, ncarly as many as received Federal

uate science students who indicated research assistantships as their-primary type of support There may be other RA recipients thatrelied on other sources for "'primary support" and are th-refore notincluded in the total in Snyder's table

"Fallows, op cit , p 6"Snyder, op cit 1981, appendix table 11, p 2t ,t nd appendix

table 17, p 270

Figure 2-10.Federal Predoctoral Fellowships am' 'Praineeships by Field,Macs! Years 1961-751

Fiscal years 1961.758aData include NIH training grant trainees In the biolcgical sciences') t not in other fields, data do not include NIMH traineesbEstimate

SOURCES For all fields except biological sciences FY 1981-1969 deli,. -ICE (1970), FY 1970-1975 data, FICE(unpublished) For biological sciencesFICE (1970, unpublished) a. d NAS, NIH trainee file, and NIH (1978)Robert G Snyder, The Effectiveness of Federal Graduate Education Policy Hod Scientific Personnel, report toOff ice of Technology Assessment, June 1985. table 2

assistance (see table 2-4). Of those students,33,000 held teaching assistantships.58

With undergraduate enrollments in both pub-lic and private institutions increasing dramati-cally, as we have seen above, the labor marketfor doctoral personnel was booming in the 1960s.________

"Graduate Student Support a,c: Manpower Resources in Grad-uate Science Educatich, l'!cF "u-40 (Wastangton, DC NationalScience Foundation, 1969), p i.)

The 260,000 new junior faculty hired during thedecade exceeded the number of doctorates awardedbetween 1960 and 1969-170,000by 53 per-cent." This situation boosted the salaries of doc-torate recipients relative to other occupations,thereby providing added inducement for more

"Allan Cartier, Ph D s and the I% .ademic Labor Mari,et (NewYork McGraw-Hill, 1976), p 123

48

people to enter doctoral study." (Faculty salariesrose 50 percent faster than those of all otherworkers during the 1960s, according to RichardFreeman.")

Thus, a complex array of interconnecting vari-ables-Federal, State, and prig to sector-appearto have combined to produce the surge of gradu-ate enrollments and doctoral production in the1960s in addition, the Vietnam War led manymale college graduates to Enroll in graduate schooland remain there through receipt of a degree.

The year 1969 turned out to be a watershed forFederal support of graduate education in scienceand engineering. A combination of political, eco-nomic, and social factors, and a growing Federalawareness that demand for doctoral scientists andengineers was not keeping up with supply, led tosevere cutbacks in Federal fellowship and trainee-ship support. The number of U.S. Governmentsupported fellows and trainees declined from36,000 in 1969 to 10,800 in 1975. The NDEA Ti-tle IV fellowship program was phased out in 1973.NSF traineeships, NIH fellowships, NASA trainee-ships, and AEC fellowships and traineeships wereall eliminated in the early 1970s. All that remainedby 1979 were 4,800 NIH traineeships, approxi-

"The emand for New Faculty in Science and Engineering (Wash-ington,DC. National Research Council, 1980), p 117

"Richard Freeman, The Market for College-Trained Manpower(Cambridge, MA Harvard University Press, 1971), pp 77-80

mately 1,500 Alcohol, Drug Abuse, and MentalHealth Administration (ADAMHA) traineeshipsand fellowships, and 1,500 NSF merit fellow-ships."

Federal R&D expenditures at universities, whichhad increased by a factor of 4 in current dollars,and a factor of 3 in constant dollars, between 1960and 1968, stopped growing. In 1974 the level ofFederal R&D support to universities was only 25percent above its total 6 years earlier in currentdollars, and 7.2 percent less in constant dollars."The number of federally supported research as-sistantships decreased by about 10 percent be-tween 1969 and 1974.

The effects of changing levels of Federal sup-port on the supply of new science and engineer-ing Ph.D.s is seen most dramatically by com-paring the experience of different science andengineering fields during the contraction period,1968-75. Figure 2-10 and table 2-5 show the dra-matic drop in the number of Federal fellowshipsand traineeships awarded in engineering, mathe-matics, and the physical sciences between 1968and 1975. The number of awards dropped from3,000 to 200 in mathematics; from 6,000 to 600in the physical sciences, and from 5,500 to 700in engineering." The total number of full-time stu-

"Snyder, op cit , 1981, pp 122 and 123"Fallows, op cit , p 4"Snyder, op int , 1981, appendix table 7. p 265

Table 2.5.- Federal Predoctoral Fellowships r..id Traineeships by Field, Fiscal Years 1961.75'

AllFiscal year sciences

Physy..,a1 3iological SocialMathematics sciences Engineering sciences sciences

1961 . . 5,445 6291962. . . . 9,245 8521963 . 11,224 8991964. . 12,757 1,0241965... 16,683 1,2111966.. 21,929 1,8241967. 928,895 2,5981968 . . 33,901 3,0241969 34,100 2,7341970 30,646 2,006

1,0731971 26,6941972 23,579 7761973 ... 19,335 4471974 13,084 3431975 .. ..... 10,787 208

1,273 638 1,762 1,123

4,940 2,3801,938

1

1,684 043,1,955

81,050

2,461 1,402 5,337 2,5332,974 3,092 6,329 3,0774,353 4,190 7,536

9,2504,026

5,424 4,784 6,8395,960 5,507 10,161 9,2495,776 5,058 10,882 9,6504,459 4,323 11,182 8,6764,071 3,540 9,669 8,3413,2121,893

2:93583 8,851 7,787

1,183 1,2258,114 6,943

4,222592 715

6,1116,098b 3,174

'Data InClude NIH training grant trainees in the biological sciences but not in other fields data do nor include NIMH traineesbEstImate

in

SOURCES For all fields except biological sciences-FY1961 69 data, FICE (1970), FY 1970 76 data FICE (unpublished) For biological sciences-ME i1970 ur,pubIlehed) and NAS NIH trainee file, and NIH (1978) Robert G Snyder The Effectiveness of Federal Graduate Education Policy (fort Sc.pritific Personnel reportto OTA. June 1985, table 2

1 ) ;

49

dents in doctorate granting departments supportedby the Federal Government in engineering, math-ematics, and physical science decreased by 18, 57,and 33 percent, respectively, between 1969 and1974 as can be seen in table 2-6.65 Figure -8 showsthe substantial decline in Ph.D.s awarded in engi-neering, mathematics, and the physical sciencesthat occurred in the period 1972-78." It shouldbe noted that those three fields were also affectedby a decline in Federal research funds awardedto universities of 20 percent in constant dollarsbetween 1968 and 1974."

By contrast, Federal fellowships and trainec-ships declined, but not nearly so precipitously,in the biological and social sciences--from 11,000to 6,000 in the former, and from 9,700 to 4,200in the latter. Moreover, as table 2-6 shows, thetotal number of full-time graduate students sup-ported by the Federal Government, including RAsand other forms of support, actually increased by45 percent in the life sciences and remained con-stant in the social sciences in the period 1969-74.In addition, Federal funding of research at univer-sities and colleges in the life and social sciencesremained at the same level, in constant dollars,throughout the 1968-74 period, as opposed to thedecline in engineering and the mathematical and

"Ib'el., appendix table 10, p. 268"Fallows, op. cit., p. 33.Snyder, op. cit., 1981, p. 262

physical sciences reported above. It is not surpris-ing, therefore, to find that the number of Ph.D.sin the life and social sciences did not decrease inthe 1970s.

The apparent correspondence between decreasesin Federal educational support in particular fieldsand decreases in the numbers of doctoral degreqiawarded several years later does not, of course,prove causality. The numbers do, however, sug-gest that there may be a relationship between thesetwo variables.

From 1974 to 1980 the Federal share of supportfor full-time science and -ngineering graduate stu-dents in doctorate granting institutions remainedrelatively constant at about to 23 to 25 percentof the total. Since 1980 it has declined is less than20 percent." Figure 2-11 presents the percentageshares of Federal support for science and engineer-ing graduate students as a whole and by field.Note the sizable variation from field to field, withmathematics and the social sciences at n percentFederal support, and engineering, physical andenvironmental sciences, and life sciences at 20 to30 percent.

"Ibid , appendix table 10, p 268, and Academic Science /Engi-neering Graduate Enrollment and Support, Fall 1982, op. cit., pp.112-113

Table 2-8.-Full-Time Graduate Students in Doctorate-Granting Institutions Supported bythe Federal Governr int by Broad Field, Selected Years

Total full-time enrollment/number federally supported 1969 1972 1974 1978 1983

All sciences 141,139 149,937 195,915 217,588 243,646Federal 51,6'.I0 45,029 48,016 51,373 47,402

Engineering 30,820 30,908 33,691 37,129 53,553ederal 12,334 11,246 10,164 10,625 11,924

Physical sciences . 30,475 27,544 29,649 31,375 35,904Federal 13,187 9,687 8,868 10,123 10,886

Mathematical sciences 11,727 11,909 13,423 13,706 19,581Federal 3,223 1,908 1,393 1,307 1,794

Life sciences 27,588 34,083 54,800 65,097 65,185Federal 11,513 12,037 16,757 19,.:59 17,221

Psychology 11,918 14,282 18,72E 20,677 21,327Federal 5,127 4,691 4,404 3,943 1,982

Social sciences 28,971 31,211 45,624 49,604 48,096Federal 6,7'4 :,460 6,4: 3 5,716 3,595

SOURCES Robert 0 Sryder Feder& Supportand National Science Foundation

of Graduate Education in tue ,.atural Sciences, doctoralA, ademic Scrence/Encneering Graduate Enrollment

diSvertatiOn Syracuse University. Jun*1981, an table 10, p 288,and Support, Fall 1983, NSF 85-300, table C14, pp 112.113

50

Figure 2.11.FullTime Graduate Students inDocforateGraraing Institutions Supported by the

Federal Government by Broad Field,Selected Years

1969 1972 1974 197E 1983

Year5.3URCES Robert G Snyder Federal Support of Graduate Education in the

Natural Sciences, Doctoral Dissertation Syracuse UniversityJune. 1981, appendix table 10. p 268 and NSF Academic Science,Engineer,ng Graduate Enrollment and Support Fell 1983 NSF85-300 table C14 pp 112-113

THE

The exact interplay of the various factors thatwill determine graduate enrollments and Ph.D.production in the sciences and engineering overthe next 20 years is difficult to predict. It wasshown above that three factors appear to influ-ence the numbers of Ph.D.s awarded in scienceand engineering fields: Federai support fur grad-uate education, State and institutional support,and demand for new doctorates. Federal supportfor graduate education in the sciences and engi-neering may rise modestly in the next decade ifthe funding of research grants to universities, and

FUTURE

hence research assistantships, continues to in-crease. Fellowship and traineeship support willprobably remain constant. State and institutionalsupport will probably decline due to declining en-rollments, with the declines being greatest in theNortheast, Middle Atlantic, and MidwesternStates. However, it is also possible that sciencear I engineering enrollments will not experiencethe overall declines of the rest of the higher edu-cation system and that States will continue to pro-vide extra support for science and engineering re-search at their State universities as part of their

51

efforts to develop high-technology industries forecor.omic growth. In that case, State and institu-tional support may not decline at all.

As for demand, we shall see in chapter 3 thatthe academic demand for new science and engi-neering faculty is likely to be very weak over thenext decade and very strong from 1995 to 2010.Industrial demand for new science and enginter-ing Phr .s, by contrast, has been growing at 7.8percent per year over the past decade and will,if it continues to grow at present rates, compen-sate for the decline in academic demand between1985 and 1995. After that, if industrial demandcontinues to grow, there will be stiff competitionbetween industry and academia for new scientificand engineering doctoral degree recipients in the1995-2010 time frame. That competition will bediscussed in chapter 3.

In the near- 'zerm, what is remarkable is theoverall flatness or the recent trend lines, since 1978

(see figures 2-12 and 2-13)," for total doctorates,for doctorates in science and engineering, and forfull time graduate enrollments in science and engi-neering. The only exception Las been engineer-ing graduate school enrollment, which has in-

"Science and Technology Resources, NSF 85-305 (Washington,DC National Science Foundation, January 1985), pp. ',"3 and 77.

35

30

25

20

15

10

5

01968 1970 1972 1974 1976 1978 1980 1982

YearSOURCE National Science Foundaticn, Science and Technology Resources,

NSF 85305, January 1985, p 77

Figure 2-12.Doctorates Awarded

Flgur 2-13.Full-Time Graduate Enrollment by Area of Science/Engineering°

Yearaln doctorate-granting Ins' ;Awls prior to 1974, data were collected from a selected group of departments rather than Porn the universebIncludas support from state and local goarnments

SOURCE National Science Foundation, Science and Technology Resources, NSF85-30D, January 1985. pp 53-65

GO

52

creased dramatically, primarily at the master'slevel. A reasonable assumption for the next dec-ade, which is that of the Scientific ManpowerCommission and the National Center for Educa-tion Statistics (see figure 2-14)," is that there willbe a continuation of current enrollment and Ph.D.production levels. Engineering is likely to continueto increase its share isnd the social sciences to de-crease theirs, due to market forces. Women,whose attainment of science s. .d engineeringPh.D.s has grown at a rate of 6 8 percent per yearsince 1973, are likely to increase their share of doc-torates in these fields, while that of men will prob-thly continue to decline. Foreign nationals willundoubtedly continue to be strongly representedamong science and engineering Ph.D.s.

The working assumption for this document isthat near-term graduate enrollments will remainconstant and Ph.D. production will remain withinabout 10 percent of its current level. In the longerrun, the strong market for science and engineer-ing Ph.D.s that should occur at the turn of the

"The Technological Marketplace (Washington, DC ScientificManp, wer Commission, May 1985), p 11, and Condition of Edu-cation, 1985, op at , pp 128-i29

Figure 2.14.Trends in Doctor's DegreesConferred, by Sex of Recipient

100,000

80,000

60,000

40,000

20,000

Total

..., 2 ...m. ...... --. .....,... "Ir

Male Female1-----......"" 1

01971 1975 1979 1983 1987 1991 1994

Year ending

SOURCE U S Department of Education National Center for EducationStatistics Condition of Education 1985, p 129

century could lead to an increase in the numberof annual doctorates awarded in those fields asstudents are attracted into them by career oppor-tunities. This development will be discussed inchapter 3 when we examine the future demandfor science and engineering doctors 1 degreeholders.

i,.."

,

Chapter 3

DemograpWc Trends and theAcademic Market for Science

and Engineering Ph.D.s

- ,1 .....,

4 .

PageNear-Term Decline 55

With Projecting the Demand forFaculty ...

of the Near-Term Trends;further Future .

Employment of Science andPh.D.s

oral Appointmentsof and Demand for

ring Faculty 80S. Education and Utilization ofForeign Scientists and Engineers 83

596465

6774

LIM of Tablesable No Page

3-1. Type of Employer of Doctoral Scientistsand Engineers (1940-82 Graduates)by Field of Ph D , 1983 70

-3-2. Type of Employer of Doctoral Scientistst and Engineers (1977-82 Graduates)

by Field of Ph.D., 198' 703-3. Selected Characteristics of Doctoral

:Icientists in Business and Industry 723-4. Comparison of 1979 Employment

Situations of Fiscal Year 1972 BiosciencePh.D. Recipients Who Took PostdoctoralAppointments v Tithin a Year After Receiptof Their Doctorates With the Situationsof Other Fiscal Year 1972 GraduatesWho Have I ;ever Held Appointments 78

3-5. Comparison ,f 1979 EmploymentSituations of Fiscal Year 1972 PhysicsPh.D. Recipients Who Took PostdoctoralAppointments Within a Year After Receiptof Their Doctorates With the Situationsof Other Fiscal Year 1972 GraduatesWho Have Never Held Appointments 79

3-6. Comparison of 10"9 EmploymentSituations of Fiscal Year 1977 ChemistryPh.D. Recipients Who Took PostdoctoralAppointments Within a Year After Receiptof Their Doctorates With the Situationsof Other Fiscal Year 1972 GraduatesWho Have Never Held Appointments 70

3-7. Origin of Foreign Students Within Fieldsof Study, 1983-84 84

List of FiguresFigure No3-1. Actual and Steady-Stae Age

Distributions Full-Time Doctoral Facultyat Ph.D. Granting Institutions, 1978

3-2. Combined Dea"--Retirement Rate,Percent of Faculty

3-3. Annual Number cf New Hires Basedoit Projections

3-4. Annual New Hires by Field3-5. New Hires by Field .-

3-6. Replacement Demand As a Percentageof the Total Academically EmployedScience and Engineering, 1983

3-7. New Ph.D.s Per 100 Academic JobOpenings . . .. .... ......

3-8. Sensitivity of Hiring Projections toVariations in Input Parameters

3-9. Annual Number of New Hires Basedon Projections . .. ... .....

3-10. Annual Hiring, Science and EngineeringFaculty, Extreme Cases . .. ... .....

3-11. Annual Hiring, Science and EngineeringFaculty, Middle Cases

3-12 Projected Annual hiring: PhysicalScientists

3-13. Projected Annual Hiring: Life Scientists .. 69 '3-14. Projected Annual Hiring: Social

Scientists . .. ......3-15. New Ph.D Hiring Science and

Engineering3-16. Possible Levels of Supply and Demand

or Physicists Within Each Five-YearPeriod, 1981-2001

3-17 Percent of Ph D Recipients From U.S.I' iversities Planning PostdoctoralStudy [aimed', tely After Receiving theDoctorate, Selected Disciplines, 1958-82 .. 77

3-18 Percent of Bioscientists PlanningPostdoctoral Study Who Had HeldAppointments Longer Than 2 to 5 Years,by Year -)f Docto-ate . .

3-19 omparison of Academic-IndustryEngineering Ph D Salaries

3-20 Total Foreign Enrollment in U.S.Institutions of Higher Education . . ... 84

3-21 Educational Attainment of Scientists andEngineers. by Immigrant Status, 1982 . 86

78

Chapter 3

Demographic Trends and the AcademicMarket for Science and Engineering Ph.D.s

THE NEAR-TERM DECLINENowhere does a confluence of demographic

trends produce more dramatic results than in theprojected hiring pattern for new Ph.D.s in the Na-tion's universitiel and colleges. Two trends are atworl. here. First, the decline in higher educationeni:Alments projected for the 1982-97 time periodby the National Center for Education Statistics(NCES) and others (see chapter 2) should reducethe total demand for faculty proportionately. Thesubsequent increase in probable enrollments overthe 1997-2010 period should lead to an increasein the size of the professoriat.

A second trend has to do with faculty retire-ments, and reelects an earlier demographic eventthe spectacular growth in college faculty thataccompanied the arrival of the post-World WarII baby boom generation on campus in the 1960sand early 1970s. The near tripling in full-timeequivalent (FTE) higher education enrollmentsthat took place between 1960 and 1975 (from 3to 8.5 million) was accompanied by an equiva-lent near tripling in the size of the full-time in-structional staff (from 154,000 to 440,000).' Thiscreated an age distribution among college anduniversity faculty in the late 1970s that was heav-ily skewed toward 35 to 40 year olds and awayfrom 50 to 65 year olds. Figure 3-1 compares theactual age distribution of facu!ty in 1978 with amodel age distribution that would be character-istic --c a steady-state equilibr'um with constantfact. Aze.2 The overrepresentation of 30 to 40year olds, and underrepresentation of 55 to cisyear olds are quite obvious.

As a result of this skewed age distribution, therate of retirement of college and university facultyis low and will continue to remain so until the

'American Council on Education, 1984-85 Fact Book on HigherEducation (New York MacMillan, 1984). rat le 114

National Research Council, Research Excellence Through the Year2000 (Washington, DC National Academy of Sciences 1979), p 19

early 1990s. Figure 3-2 shows the combined deathand retirement rates projected for the majorscience and engineering fields through the courseof the century by Charlotte Kuh.3 A ". Lady state"re., assuming all faculty work for 35 years, fromre& ipt of Ph.D. at age JO to retirement at age 65,and no faculty growth, would be 2.5 to 3.0 per-cent per year.' It is clear that none of th fieldsdisplayed in figure 3-2 reach that rate !Trail the1990s, and only the social sciences surpass 2.5 per-cent per year before 1994. Thus the per;od of lowfaculty retirement and death rates coincides withthe period of possibly declining student enrollments.

The confluence of these two trends leads to aperiod of extremely weak demand for new doc-torates, ir academia from the early 1980s throughthe mid-1990.. Figure 3-3 illustrates the gap be-tween the projected demand for new fa -ulty,which averages 6,500 to 16,000 per year duringthis periods and annual awards of doctoral de-grees, which average about 30,000 per year. Thiscompares quite unfavorably with the 1960s when,as shown in chapter 2, the supply of new Ph.D.saveraged 17,000 per year and the demand forfaculty (including both doctoral and nondocroralfaculty) averaged 26,000 per year. It is also con-siderably worse than the situation in the early1970s, when the supply of new Ph.D.s averaged31,000 per year and the demand for faculty(growth plus replacements) also averaged about31,000 per year, according to OTA analysis!' It

'Ibid , p 23'Ibid , p 24'William G Bowen, Report of the President (Princeton, NJ I'nnce-

ton University, April 1981), p 20. Charlotte Kuh and Roy Radner,Preserving a Lost Generation, report to the Carnegie Council onPolicy Studies in Higher Education, October 1978. p 9

'According to the National Center for Education Statistics therewere 403,700 FTE faculty employed in 1970, 502,700 in 1975 ar,c1528,700 in 1980 That translates to growth rates of 19,800 per yearbetween 1970 3nd 1975 and 5,200 per year between 1975 and 1980;U S Department of Education, Condition of Education (Washing-

((otitimted on p 571

55

4

58

Figure 3.1. Actual and Steady-State Age Distributions,FullTime Doctoral Faculty at Ph.D.Granting Institutie4s, 1978

5

4

iIs

I2

I

030 40 50 00 70

A04

SOURCE National Research Council, Research Excellence Through the Year 2000 (Washington, DC National Academy Press, 1979), p 19

C,

t

Figure 3-2.Combined DesthRetirement Rate, Percert of Faculty

1976 1980 1964 1966

Year

1992 1996

EDLife 'Gloom El Social sciences 0 Physical sciences le Mathematics III Engineering

SOURCE NatIcnal Research Counch, Research Excellence Through the Year 2000 (Washington DC National Academy i ess 1979) p 23

-t f 0

IIMME11!.

35,000

30,000

25,000

20,000

15,000

10,000

5,000

Egure 3.3.Annual Nt.mber of New Hires Based on Projections

Ph D s awardeci,yr (1983)

31,000

57

1970-75 1975-80 1980-85 1985-90

Five-year intervals

0 Princeton model Kuh-Radner model it*

9,800

6,800

1990-95 1995-00

OTA calculated from historical data

SOURCES William G Bowen, Report of the President )Princeton, NJ Princeton University, April 1981) p 20, and C Kuh and Rreport to the Carnegie Council on Policy Studies in Higher Education October 1978, p 40

is roughly comparable to the late 1970s, when31,500 doctorates were awarded annually, but thenumber of new faculty slots averaged approxi-mately 19,000 per year. Although over 40 per-cent of all science and engineering Ph.D.s enterindustry or government, it is important to remem-ber that most analysts assume that no more than----- ---(continued from p 55)

ton, DC: U.S. Government Printing Office, 1985), table 2 12, p.110. According to the National Research Council the combined deathand retirems.nt rule for faculty in the 1976a was about 1 2 percentper year; National Research Council, op. cit p 23. This translatesinto total deaths and retirements of 4,200 per year between 1970and 1975 and 6,000 per year between 1975 and 1980. Charlotte Kuhand Roy Radner .-eport tenured and nontenured quit rates of 0.5and 4 percent per year in the mid-1970s; Kuh and Radner, op. cit ,

table A-4, p. 41. Using those quit rates and the ratio of tenured tonontenured faculty reported by Kuh and Radner in the same source(70:30) resignations can be estimated to be 6,300 per year from 1970to 1975 and 7,800 per year from 1975 to 1980. Adding growth plusretirement plus resignations for each 5-year period gives the totalannual demand for new faculty of 31,000 for 1970-75 and 19,000for 1975-80 reported in the text

Radner, Preserving a Lost Generation.

50 or 60 percent of all new faculty positions willbe awarded to holders of the doctoral degree. Theremainder will go to M.A.s, M.D.s, J.D.s, D.D.S.s,and other first professional degree holders.

This decline in academic positions available tonew Ph.D.s was first projected in the early 1970sby Alan Canter. Ills projections were refined andmodified by a number of analysts, including RoyRadner, Charlotte Kuh, and Louis Fernandez' be-Meen 1976 and 1980, but the basic findings re-mained the same. Kuh and Radner were so dis-turbed by the impPcations of this finding for thecareer- "f young scholars that they entitled their

'Alan Cartter, Ph.D.s and the Academic Labor Market (NewYork: McGraw 1976); Louis Fernandez, U.S. Doctorate FacultyAfter the Bocm: Demographic Projecaons to 1995, Technical Re-port No. 4, Carnegie Council on Policy Studies in Higher Educa-tion, October 1978; Kuh and Radner, op. cit.; Roy Radner and L.S. Miller, Demand and Supply in U.S. Higher Education, a reportto the Carnegie Council on Policy Studies in Higher Education, 1976.

t

58

1

I

11111

10,000

8,000

6,000

4,000

2,000

Figure 3.4. Annual New Hires by Field

Ph D.s awarded/yr (1983)

11978 1980

Life sciences 0 Social sciences

I1984

III

Year

1988 992 1998

SOURCE National Research Council, The Demand for New Face, , a, Science and Engineering. 1979 (Washington, DC National Academy Press. 1980), pp 178 and 177

study for the Carnegie Council on Policy Studiesin Higher Education, Preserving a Lost Genera-tion: Policies to Assure a Steady Flow of YoungScholars Until the Year 2000.

In 1978 the National Research Council (NRC)asked Kuh and Radner to adapt their forecasts toscience and engineering fields, as part of a studyon Research Excellence Through the Year 2000(reference 2, above). Using National Science Foun-dation (NSF) and NRC data, Kuh and Radner pre-pared a set of projections for the number of newfaculty hires that would be required each yearfrom 1978 to 2000 for science and engineering asa whole, and for the major broad field catego-ries. Those projections are summarized in figures

3-4 and 3-5 which also compare the number ofnew hires with the number of doctoral degreesawarded in 1983.° As can be seen, Kuh and Rad-ner predict a weak academic market for scienceand engineering Ph.D.s from the early 1980s tothe early 1990s followed by sustained growththrough the year 2000. (The apparent growth innew hires in the physical sciences between 1980and 1988 is the artifact of a spuriously high non-tenured quit rate built into the model for thatgroup.)

'National Research Counci. The Demand for New Faculty inScience and Engineering, 1979 (vVashingtor DC National Acad-emy of Sciences, 1980), pp. 62-63, 176-177

59

Figure 3-5.New Hires by Field

Year

IIIMathematical sciences El Physical sciences . EngineeringSOURCE National Research Council, The Demand for New Faculty in Science and Engineering, 1979 (Washington, DC National Academy Press, 1950), pp 175 and 177.

PROBLEMS WITH PROJECTING THE DEMAND FOR NEW FACULTYTo detemine whether or not Kuh and Radner's

projections of a hiring decline from 1983 to 1995are correct it would be useful to compare theirnear-term (1975-83) projections with recent his-torical experience. Unfortunately, that is a verydifficult task, for two reasons. First, and most seri-ous, is the extraordinary fact that no organiza-tionnot NSF, nor NRC, nor NCESactuallymeasures the number of new faculty appointedin a given year. Thus we cannot compare projec-tions with any directly observed data.'

'In its most recent survey of Ph.D.-granting science and engineer-ing departments the National Science Foundation asked chairmento estimate the number of new full-time permanent appointmentsmade in 1983. Unfortunately, this is a limited sample, and the re-

Peter Syverson and Lorna Foster of NRC haveattempted to infer hiring rates indirectly fromNRC's biannual Survey of Doctoral Recipients(SDRs), a longitudinal study of the career patternsof a 10-percent sample of science, engineering, andhumanities Ph.D.s. Comparing responses to the1981 and 1983 SDRs, Syverson and Foster esti-mate the growth in the number of faculty posi-tions as "the change in the number of academi-cally employed respondents between 1981 and1983" and attrition as "the proportion of respond-

sults have not yet been published. Moreover, NSF has some reser-vations about the accuracy of reporting or. this question. Personalcommunication from Christine Wise, Science Resource Studies Di-vision, National Science Foundation, Sept. 23, 1985.

80

ents who no longer reported academic employ-ment" in 1983.10 Syverson and Foster find that"evrn when the academic labo market is in anoverall steady state, as it now appears to be, therecontinues to be demand on the order of 5 to 7 per-cent" for both replacement and growtS. See fig-ure 3-6 for the percentages by field. Comparingannual doctoral degree production for 1983 withthe calculated number of job openings they findratios ranging from about 190 new Ph.D.s per 100academic job openings in engineering and physi-cal science to 109 per 100 in mathematics andcomputer science, with the life and social sciencesdirectly in the middle." (See figure 3-7.',

For individual fields we can compare Syversonand Foster's calculated new hires with Kuh andRadner's projected new hires for 1983. The resultsare displayed below.

Although the totals for science and engineeringare reasonably close, there are clearly serious dis-crepancies between the projections and the cal-culated new hires in all fields, with the greatestdifferences appearing in the social sciences andengineering. The fact that some fields are substan-tially overprojected while others are underpro-jected indicates that the problem is probably notsimply one of a lack of comparability in the pop-ulations being studied.

A second factor limiting the ability to treat pro-jections from models such as that of Kuh and Rad-ner as precise numerical predictions is the extremesensitivity of those projections to assumption?made by the modellers about certain key param-eters connecting enrollment to the demand for

"Peter Syverson and Lorna Foster, "New Ph.D.s and the Aca-demic Labor Market," Office of Science and Engineering PersonnelStaff Paper No. 1, typescript, p. 2.

"Ibid., pp. 3 and 10.

Fig. a 3-8.Replacement Demand As aPercentage of the Total Academically

Employed Science and Engineering, 19838

7

6

5

2

i

Physical science30 273

El Growth

Life scisnce48,588

Math/computer science Social scie13,105 53,449

L.S.16._

Attrition

Engineering18,382

SOURCE Peter Syverson and Lorna Foster, NOW Ph D a and the Ac 4emicLabor Millet (Washington, DC National Research Cc _nci, .9134),pp 4 sh.. .

faculty. These parameters include retirementrates, tenure and przmotion rates, voluntary "quitrates," and variations in the overall ratio of facultyto students from year to year. Robert Klitgaardreports that "reasonable variations" in one param-eter alone, "quit rates," lead to projections of newhires differing by a factor of 5 or 6 by 1985.12

"Robert E. Klitgaard, The Decline of the Best? Discussion PaperNo. 65 D (Cambridge, MA: John F. Kennedy School of Govern-ment, Harvard University, May 1c79), p. 16.

New Academic Hires, by Science and Engineering Field, 1°03

Ci 0

51

150

100

50

0Engineering

Figure 3.7. New Ph.D.s Per 100 AcademicJob Openings

Science and Engineering, 1983

NO.

192 191

j/

200

150

10t.

50

0

Physicalscience

1584Tv

/2./

Ufescience

149

,

SocialAclence

Humanities, 1983

109

777,

Math/computerscience

154. 118 112 111WAN MEM wows.

Wvr/, / 1

,

..,

,

Other Music Foreign History Englishhuman L & L !mer.

SOURCE Peter Syverson and Lorna Foster, N. oh D s and the Academic-Ow Market (Washington, OC National Research Council, 1984),

4 and 7

Lee Hansen and Karen Holden have calculatedthe effect of changing the mandatory retirementage from 65 to 70. They find that a 5-year shiftin the retirement age assumed 'n their projectionmodel causes a 63-percent drop in the projectednumber of new faculty hired in the 1987, and a23- percent drop in projected new faculty hires in1992.'3 For 1987, the number of projected newhires falls from 38.8 percent of its 1977 level to14.5 percent, while for 1992 the projected percent-age declines from 44.2 to 33.9 percent.

Klitgaard, in a provocative essay on the pitfallsof forecasting academic demand, shows the effects

"W. Lee Hansen and Karen C Holden, "Critical LinkamHigher Education: Age Compositicn and Labor Costs," InsuranceMathematics tic Economics, vol. 4, No. 1, January 1985, pp. 60-61.

of changing assumptions about retirements, "quits,"promotions, and faculty-student ratios. Takinga model de Aoped by Louis Fernandez, Klitgaardsimultaneously varies each of the above-men-tioned parameters to the limit of its range of un-certainty and then adds or multiplies the variancesproduced. The result is depicted in figure 3-8which shows a variation of a factor of 6 betweenthe sum (or product) of the extreme highs and thesum (or product) of the extreme lows of all of thevariables taken together."

In April 1985, at the request of OTA, NRC con-vened a panel of academic labor market specialiststo determine whether the projections made in thelate 1970s are still valid, in the light of new dataand subsequent analyses. The panel devoted agreat deal of attention to the many uncertaintiesthat surround projections of student enrollments,faculty retirements, "quit rates," promotions, ten-ure decisions, and salaries. However, there wasgeneral agreement that the overall direction ofchange projected in the late 1970s was still valid.According to Michael McPherson, a participantat oth the 1978 and 1985 NRC workshops:"

0- pent that is very much worth underlin-ing is thac we all seem to be in rough agreement:the 1970s projections of overall demand forter -hers for the next decade aren't obviouslywrong. People knew there were uncertaintieswhen they were made; of course, uncertaintiesstill exist, but if there is something wrongw* the projections, it is not anything tnat wehave detected . . . .

. . . . by and large, all people looking at thisquestion generally have the same picture of whatwill occur. There is much agreement both aboutthe major quantitative factors impinging on thisset of labor markets and, at a broad level, abouthow the market will react to those forces.

There was also substantial agreement that de-spite the consensus c overall trends, the rangeof uncertainty in the factors that influence demandmake any numerical comparisons with predicted

"Klitgaard, op. cit., p. 19."National Research Council, "Draft Proceedings of the Work-

shop on the Forecasting of Demand for University Scientists andEngineers," Apr, 8, 1985, pp 32 and 49.

70

02

Figure 34.Sensitivity of Hiring Projections to Varations in input Parameters

,,,. High multiplicative-- High ac:::!!t:qe

4.Fernandez baseline

..- Low multiplicativeLow additive

18,0001

15,000

1 12,000

1iIII,T 9,000

iIi 8,000

3,000

41\

%

/ II

II ///I\-Vol /

I\A

II

0 I1978 1980

Year

i1992 1996 2n00

Nate The variability represented hue does not irclude uncertainty In enrollments Calculations and graph by authorSOURCE: Robert E. Klitgaerd. The Doan, of the Boof? Discussion Paper No 65D (Cambridge, MA John F Kennedy School of Government, Harvard

University. May 1979), p. 19.

supply very problematic. McPherson summarizedthese uncertainties as follows:"

The workshop discussions . . . . highlighted anumbis of complications and qt.alifications, someof which were acknowledged in the earlier work.Ar.vag the most iwtportant of these were the fol-lowing:

1. The Research Excellence study focused onthe demand for teaching faculty, but re-

"Wad S. McPherson, "Numbers and Quality: Anyzaas theMarkt for University Sdentists and Engineers," National ResearchCouncil, Pommeling of Demand for University Sdentists and Engi-am* Pnamedfnip of a Worbhop (Wathinipon, DC: National Acad-emy Press, 103), p. 1.

search support generates academic hiringtoo, especially in major universities. Someuniversities are evolving elaborate patternsof nonfaculty research staffing, partly in re-spond to anticip..,ed declines in teachingpositions. To the extent that universities cansuccessfully decouple teaching from researchhiring in t'.:1.4 way, the effects of fluctuationsin enrollment on research effectiveness maybe mitigated.

2. Research Excellence and related studiestreated the pace o; retirements . .; mechani-cally determined by faculty demographics,whereas in fact retirement should be seen asa decision influenced by, among other

63

things, the economic incentives facing re-tirees.

3. Research Excellence followed the modelingwork of Radner and Kuh in supposing thatuniversities would respond to declining de-mand for faculty by reducing the rates atwhich they promoted people to tenure . . . .

So far that is apparently not happening:universities are hiring more faculty off thetenure rick but continue to promote tenure-track faculty as before.

4. The models on which Research Excellencerelied dealt with academia as a whole, butin fact different segments of the academicsystem may behave very differently in theyears ahead.

5. It is similarly important to recognize thatdifferent fields within science and engineer-ing are likely to fare very differently, owingboth to variations in research funding andin student course preferences.

Because of these uncertainties, participar- e. atthe wcrkshop generally felt that it is more impor-tant to analyze existing data and refine our under-standint If how the academic labor market worksthan to make detailed projections of the numbersof Ph.D.s required or produced in specific fields.In the words of Stephen Dresch:17

. . . jwe need' to have SOfile ita5Ortably sus-tained work in this area . . . nobody systemati-cally asks, "What data should be collected to an-swer interesting questions?" Therefore, a lot ofmoney used to collet data is, in fact, wasted be-cause the data collected has fatal flaws: it trackssome flows in the system but doesn't permit track-ing other flows, a situation that could very eas-ily have been rectified if one had been striving fora complete picture. A lot of data, in fact, is in-complete and can't be spliced together. The otherside is that, for O. practical purposes, we reallydon't want to pay anyone to learn anything withthese data. What we essentially want is to "buy"the aura of rationality in action . . .

McPhe. son summarized the situation as follows:"

On the qualitative side, we really need to un-derstand better how this system behaves, recog-nizing that institutions and individuals adjustwhen conditions change. The simplest kinds ofmanpower forecasting models assume a great dealof rigidity That people in institutions just con-

"Natiotial Research Council, "Draft Proceedings," op. cit , p. 29."Ibid., p. 51.

tinue to do what they used to do in the face ofrac%cal changes in conditionsbut we know thatis false, because in one way or another gaps getfilled . . . . It is in the adjur .ment process that bet-ter qualitative understanding is needed. For thatpurpose the quantitative models are not really somuch intended to be accurate predictions of thefuture as guidelines that shove us where the ad-justments will have to occur and direct our un-derstanding.

No- v, to achieve this qualitative understanding,we don't really want co develop a super-elaborate,sophisticated structural model . . . . Instead, weshould study the many variableswage behavior,tenure policies and how they respond, how uni-versities handle nontaculty research positionsand build up from that a better understanding,not so much to predict but to understand how thisbeast operates. What we need is more basic re-search into these labor markets.

Participants at the workshop listed the follow-ipg items for research on the qualitative aspectsof the academic Sabo- market:"

1. How is the quai,ty of students attracted toadvanced study imque:iced by changes in la-bor market conditions? Will declining de-mand cause the best students differentiallyto select themselves out of scientific careers?

2. How do academic departments cope withhiring shortages like those now being ex-perienced in engineering, and with what im-plications for research and teaching effec-tiveness?

3. How, and how effectively, do universitiesrespond to reductions in demand for teach-ing faculty?

4. Whg.t factors influence the mobility of expe-rienced Ph.D.s, both among fields and betweenacademic and nonacademic employment?

Peter Syverson of NRC summariz.1 the stateof the art of forecasting supply and demand inthe academic market as follows:20

At present out ability to forecast changes in theacademic ma4'et is decidedly limited. For any-thing more Timed than seat-of-the-pants plan-ning, a far more precise model must be developedto accurately reflect the academic market up toand beyond the turn of the century.

"McPherson, op. at , p 4"Syverson, op cit., p 12

4

84

IMPLICATIONS OF THE NEAR-TERM TRENDSThe implications of the overall trends projected

by labor market specialists in the late 1970s, andconfirmed by the NRC workshop in April 1985,were discussed at great length in conjunction withthe early forecast of a declining academic labormarket for Ph.D.s in the 1980s and early 1990s.Kuh and Radner, for example, in their work onPreserving a Lost Generation argued that:"

It is in 'ooth the national interest and the inter-est of individual institutions to assure a moder-ate but steady flow of young doctorate scholarsinto academia . . . . An academic enterprise inwhich half as many young faculty did twice aEmuch teaching could not help but result in a con-siderably smaller amount of research, with con-siderable consequence for U.S. science . . . .

When fewer and fewer people can be hired, thepredictors (of creative and lasting scholarship) arelikely to become more and more conservative.The young Ph.D. who has two published articlesin addition to his thesis is likely to be chosen overthe young Ph.D. who has an interesting area ofresearch with a longer gestation period. "Mis-takes," after all, are much more costly when theycan be spread over fewer people. But, in fact, theresearch with the longer gestation period may bemore productive in the 1( 4 run . . . . Prograincare needed which allow (the research universities]to take some long shots" in the hiring of youngscholars. The larger the pool, the more likely thatthe best scholars will be found in it.

The 1979 National Academy of Sciences reporton Research Excellence Through the Year 2000:The Importance of Maintaining a Flow of NewFaculty Into Academic Research, cited above,enumerated in some detail the possible conse-quences for science of the projected decline in aca-demic hiring of Ph.D.s. According to t'ie Com-mittee, "damage to the nation's research effort islikely to result from the expected constriction inthe flow of new faculty" for a nember of rea-sons:"

1. the rate of research innovation, the inflowof new ideas, and the vitality of the researchenvironment will be impaired;

"Kuh .:nd Radner, op cit , pp 2-3"Nationai RiNearrh Council, Rese..rch Excellence op cit p vi

2. continuity in the education and socizlizatioaof succeeding generations of researchers willbe threatened; and

3. the perceived lack of opportunities for anacademic career may discourage able andcreative young people from pursuing careersin basic scientific research.

The Carnegie Council on Policy Studies inHigher Education discussed the "implication of thedemographic depression for faculty" in its 1980report on the future of higher education:23

The tenured professoriate in 4-year colleges willkeep on aging with the ages of the modal grouprising from 36 to 45 in 1980 to 56 to 65 in 2000.This will increase the age gap between studentsand faculty, raise the average cost of faculty sal-aries, and make it hard to in! -aduce new fields,new courses, new subject matter. Tenure ratios,wl...ch were 50 percent as recently as 1969, havenow risen to 75 percent; and colleges encounternew rigidities in redeploying their resources.

Some faculty members are potentially muchmore affected than are others: In the East andNorth more than in the South and West; in com-prehensive colleges more than in community col-

irt !fss -aelectiv..:vie than flan e selec-tive 4-year liberal arts colleges; in doctorategranting universities more than in research univer-sities.

The response to these cries c i alarm from dis-tinguished educators and analysts is summarizedin a review of the current state of the academiclabor market in late 1984 by Michael McPherson.He finds that, to a remarkable degree, "the pol-icy issues identified in the late 1970s and earl-,19q0s" relative to the declining academic marketfor Ph.D.s "remain central and largely unresolvedtoday. "24

'Three Thousand Futures The Next Twenty Years for HigherEducation (San Francp.(0, CA ;woe)/ Bass, 19130), pp 80-82

"Michael S McPherson, 'The State of Academic Labor Markets,-November 10,4 draft, p 5, to be published iiBLR Smit (edThe State of Graduate Education (Washington, DC The BrookingsInstitution, 1Q851

65

THE FURTHER FUTURE

All of the studies of the late 1970s ended theirprojections at the year 2000. Unfortunately thatis just the beginning of a pronounced "rebound"period, when enrollments are expected to increaseand large numbers of faculty hired in the 1960sand early 1970s are expected to retire. McPher-son expresses the conventional wisdom when hestates that "to the degree that student and facultydemographics are determining factors, data avail-able now suggest that recovery will come quitelate in this centui y and will not be strong."25 How-ever, projections for the first decade of the 21stcentury make the transition to growth appear con-siderably less smooth than McPherson argues. Toexamine the "rebound" in greater depth, OTA per-formed its own analysis of the period 2000-10,

"Ibid., p. 29.

3

35,000

25,000

20,000

15,000

:0,000

5,000

using the model developed by Herring and San-derson for the 1981 Report of the President ofPrinceton University cited above. OTA first ex-tended the Princeton model to the year 2015, usingthe same values for the input parameters as Her-ring and Sanderson, and population trends for2000 to 2015 from the Census Bureau Projectionsof the Population, middle series, cited in chapter2. The results, displayed as the white bars onfigure 3-9, show a tripling in demand for newfaculty, from 6,500 per year between 1980 and1995 to 20,000 per year between 1995 and 2010.

OTA then revised the Princeton model to in-corporate a different set of assumptions aboutresignations and retirements, which OTA believestrack recent data on these two phenomena moreclosely than those of Herring and Sanderson. The

Figure 3.9. Annual Number of New Hires Baud OR PEOiroalOna

Ph D s awarded/yr (1983)

Five-year intervals

Princeton model CI Revised modelSOURCES Milan G Bowen. Report of the President (Princeton. NJ '2ronceton University April 1981). and Office of Technology Assessment analysis

66

results are displayed as the striped bars or figure3-9. As can be seen, the OTA assumptions leadto a level of new faculty hiring in the 1980-95 timeframe that is nearly double that of Herriig andSanderson-12,000 per year. In the first uecadeof the 21st century, if tl-e assumptions in the OTAsimulation hold, the number of new academichires per year could reach 25,000 per year, or dou-ble that of the early 1990s.

Hansen and Holden project that new facultyhires could increas by a factor of 4 between 1987levels and 2002. In their worst case sceLtari, inwhich most faculty choose to retire at age 7, theincrease between 1987 and 2002 would be a fac-tor of 10.26

The implications of this dramatic reversal offortunes for universities and new science and engi-neering Ph.D.s have not been examined to date.Although it appears that the academic demandfor new faculty will remain below the annual sup-ply of new Ph.D.s, even in the boom years 2000to 2010, that situation could, in fact, change dra-matically if graduate school enrollments dropsharply in the 1990s due to the weak market fornew faculty in that time period. In other words,if graduate students respond to the market sig-nals of the ea-:y 19-73s by decreasing their enfoll-ments in loral programs, there could be toofew of the- . to meet the surging demand that illoccur at the turn of the century. To avoid th.,potential market failure, it would be prudent tomonitor new academic hires and, if the trends dis-cussed above materialize, possibly institute coun-tercyclical support programs in the early 1990s.

Even if the supply of new Ph.D.s increases tomeet demand in the early 21st century, a coun-tercyclical policy for the early 1'40smay be worthconsidering for a second reason. The large num-ber of new faculty hired ir. the 1995-2010 timeperiod will produce an age distribution amongfaculty in the second decade of the 21st ceiturythat is heavily skewed toward the young, as wasthe age distribution in the late 1970s. (See figure3-1, above.) This could lead to a repeat of the lowretirement, low replacement situation of the 1980-95 time period. To avoid a repetition of the 'boomand bust" cycle of the 1960s and late 1970s it may

"Hansen an.; Holden, op. at., p. 60.

be worthwhile to attempt to formulate a coun-tercyclical policy to stimulate demand in the early1990s.

The science and engineering fields appear to )l-low the g ^era', academic market fairly closely.OTA appli.d a revised version of the Princetonmodel to the most recent data from NSF on theage distribution of doctoral scientists and engi-neers employed at 4-year colleges and universi-ties.27 Using 1983 as a base year, OTA is ab'-to project the number of new junior acule..,appointments that would be made at 5-year in-tervals from 1983 to 2013, under different assump-tions about retirements, resignations, and enrc,11-ments in science and engineering courses. Figure3-10 shows the variation in the demand for newscience and engineering faculty over the 30-yearperiod under two extremc assumptions: one ofhigh demand for science and engineering faculty,high rates of retirement among ser.ior faculty, andhigh rates of 7--ignation among junior faculty;and one of low demand, low retirements, and lowresignations. These appear to re -resent reason-able est' 5 of the extreme levels of possible aca-demi' ng of new science and engineeringPh.D.. ver the next three decades.

To refine its analysis, DTA app:ied its modelto historical data on Ph.D. scientists and engineersin educational institutions for the years 1973 to1983 published by NSF." OTA found that a highdemand, low retirement, low resignation scenariobest fit historic data. Two versions of such a sce-nario are displayed in figure 3.11 which representsan educated guess based on historical experienceas to the likely range of academic hiring of scienceand engineering Ph.D.s over the next 30 years.The number of science and engineering doctoratesappointed by colleges and universities in the 1977-82 time frame, as calculated by OTA from NSFdata, is also displayed in figure 3-11 as an av -age number of new hires per year over the 5-yearperiod. As can be seen, the number of new scien-tists and engineers appointed by colleges anduniversities declines significantly in the 1983-98time frame. It then increases to more than 9,000per year in the 1998-2013 time frame. Similar pro-

"Charactenstics of Coctoral Scientists and Engineers in the UnitedStates. 1983, t: '..+F 85-303 (Washington, DC. National Science Foun-dation, 1985), table B-5, pp 19-23

"Ibid., table 3, pp. x and xi

jections for individual fields follow similar pat-terns, as can be seen from figures 3-12, 3-13, and3-14 which show the hiring trends for the physi-

15,000

12,000

9,000

8,000

67

cal sciences, the life sciences, and the socialsciences under the two scenarios most consistentwith past experience.

Figure 3-10.Annual Hiring, Science and Engineering Faculty, Extreme Caen

M=0

3,000k

01983-88 1908-91

Low assumption III High assumption

SOURCE Office of Technoiogy Assessment

1993-98 1998 -03 2003-08 2008-13

Ave-year intervals

INDUSTRIAL EMPLOYMENT OF SCIENCE AND ENGINEERING PH.D.s

The trends discussed above in the academic em-ployment of science and engineering Ph.D.s mustbe placed in the context of the availability of alter-native careers for those professionals. As tables3-1 and 3-2 show, educational institutions employonly 52 percent of all Ph.D. scientists and engi-neers, and less than 48 percent of those who re-ceived their doctorate after 1977. tie variationfrom field to field is substantial, with mathe-mat!cs, social science, and biology havin" two-

thirds or more of their Ph.D.s employed in acade-mia, while chemistry and engineering are at one-third or less academic employment. By compari-son, in the humanities 83 percent of all Ph.D.r,are employed at educational institutions." Ofcourse, the principal alternative market for scienceand engineering Ph.D.s is industry, which has in-

"Nat:oral Research Council, Science, Engineering and Humani-ties Doctorates in the U S . 1983 Profile (Washington. DC NatioaalAcademy Press, 1983), pp 36, 37, and 74

U

68

Figure 3-11.Annual Hiring, Science and Engineert-tg Faculty, Middle Cases

15,000

12,000

1968-93 1993-96 1998-03 2003-08 2006-13

Five -year intervals

0 Historic assumption 13:1 1982 hiring Level Middle assumption

SOURCE Office of Technolog-' Assessment

Figure 3.12. Projected Annual Hiring: Physical Scientists

Historic assumption WOOD assumption

SOUR(,E Office of Technology Assessment

Five-year intervals

I

69

Figure 3-13.Projected Annual Hiring: Life Scientists

1911346 19118-93 1903-98 1908-03

0 Historic assumption 131 Midday assumption

SOURCE Office of Technology Assessment

ita

-6

1Iz

Palest Intervals

2003-011

Figure 3-14. -Projected Annual Hiring: Social Scientists

2008-13

1983-88 1111111-93 103-1111 19015-03

Finless Intervals

0 ilistorIc assumption Ei Middle assumption

SOURCE Office of Technology Assessment

4 (j

2003-0B 20(111-13

Table 3.1. Type of Employer of Doctoral Scientists and Enginaers (194042 Graduates) by Field of Ph.D., 198.5 (In percent)

Type of employer

Field of doctorate

All fields MathematicsComputer

sciencePhysics/

astronomy ChemistryEarth/

environment Engineering Agriculture Medicine Biology PsychologySocial

scienceEmployed population' 350,900 18,500 2,500 28,100 44,900 12,000 55,000 16,900 12,900 57,600 47,500 55,000Educational institutions 52.2 73.0 46 5 47 5 32 6 45.4 34 9 57 8 56.5 64 1 47 1 71 8Business/industry° . 31.7 19 2 47 ( 36.3 57.3 30.0 54 2 22 4 24 2 18 0 27 5 12 8U.S. Government. . 7.6 5 1 4 2 10.8 5 8 17 9 7 1 14 0 7 3 8 7 3 7 7 2State/local gc ment 2 0 0.2 0 1 0 3 0 7 3.3 0 6 2 5 2 8 2 1 4 7 3 0Hospital/clinic 2 8 0 3 0 9 0 9 0 1 0 2 0 2 5 8 3 2 12 7 0 4Other nonprofitorganization 3.1 1 8 1.4 4 1 2 2 2 9 2 6 2 2 3 1 3 6 4 1 3 3Other .. 03 01 01 05 01 02 13No report . 0.2 04 01 02 03 0.4 02 04 04 02 0,_ 02headadas iad and part -lone emploped ones

%Weeps low-straNesee

SOURCE National Ronan* Council 1983 Pro*. ro 36 37

Table 3.2. Type of Employer of Doctoral Scientists and Engineers (197742 Graduates) by Field of Ph.D., 1983 (In percent)

Type of employer

Field of doc '^

All fields MathematicsComputer

sciencePhysics/

astronomyEarth/

Chemistry environment Engineering Agriculture Medicine Biology PsychologySocial

scienceEmployed population' 80,800 3,600 1,200 4,300 7,000 3,000 11,500 3,800 3,800 11,100 15,900 15,600Educational institutions 47 8 70 7 51 8 35 / lb 4 42 8 32.1 55 7 55 3 59 2 42 3 64 8Bitatess/industryb 34.1 21.6 44 3 44 0 73 6 35.3 57 2 29 1 23 2 22 8 27 2 17 1U.S. Government 66 58 30 132 38 148 58 86 78 77 32 78State/local government 3 1 0 1 0 9 3 5 1 0 2 3 2 2 2 7 6 1 4 9Hospital/clink . 4 3 0 8 0 9 1 1 0 5 7 3 5 0 16 4 0 6Other nonprofit

organization 3 7 i 1 0 9 5 7 2 2 3.0 3 4 2 5 3 7 4 5 4 4 4 3Other 0 1 0 4 0 9 0 4No report 02 06 02 09 04 02 04 0 ?ofoupdva WM and Parl-unts limPloYEI ernbinduels sell-empleyerl

SOURCE National Research Cane 1983 Proble pp 36 37

71

creased its share of the employed doctorates inthose fields by 10 percent in recent years. Inphysics, chemistry, and biology the industrialshare of employed doctoral scientists is 20 per-cent higher among recent Ph.D.s than it is in theoverall population of doctoral degree holders

The increasing share of science and engineer-ing Ph.D.s employed in industry is an importantphenomenon that deserves closer scrutiny. In1973, there were 53,400 Ph.D. scientists and engi-neers employed in industry, as compared to 129,300in academia, according to tne National ScienceFoundation.3° By 1983 the number of industrialscientists and engineers had more than doubled,to 113,500, while the number in academia had in-creased by only 50 percent, to 196,100. The rateof growth in industry was about 8 percent per yearover the decade while that of academia was morelike 4 percent per year.

Among the different science and engineeringfields, computer scientists, psychologists, and so-cial scientists showed the most dramatic growthin industry between 1973 and 1983, increasing by3:0 to 500 percent (see cable 3-3).310f course, allthree fields started from very low levels in 1973-3,000 or less. Engineering and physical sciencecontinued to employ the largest numbers of doc-ti.:ral scientists and engineers-34,500 and 29,0Wrespectively in 1983but their grov th rates overthe decade were lowest at 100 and 46 perccnt re-spectively. The number of industrial Ph.D. scien-tists and engineers whose primary work activitywas research and development doubled over thedecade, but the numbers in consulting, sales andprofessional services, and "other" all quadrupledor quintupled. Fifty percent of the growth overthe decade was in these three non-R&D relatedactivities.

Even more dramatic, however, was the changein the relative demand for new Ph.D. scientistsand engineers between industry and academia. Be-tween 1973 and 1975 the number of academicPh.D. scientists and engineers increased by 19,700,while the number of industrial Ph.D. scientists and

"Science and Engineering Personnel: A National Overview, NSF85-302 (Washington, DC: Nation; Science Foundation, 1985), ta-ble B-12a, p. 113.

"Characteristics of Doctoral Scientists and Engineers in the UnitedStates: 1983, op. cit., table 4, p. xii-xiii.

V

engineers increased by only 11,200. Between 1981and 1983, by contrast, the corresponding increaseswere 9,000 for academia and 14,300 for indus-try.32 Thus the demand for additional science andengineering Ph.D.s in industry increased from 55percent of the academic demand in the early1970s, to 160 percent of the academic demand bythe early 1980s.

To get a complete picture of the academic andindustrial markets for science and engineeringPh.D.s we would need to add the replacement ?-.-mand to the growth demands calculated above.The replacement demand in academia can be eas-ily calculated with the OTA model describedabove, and it comes to 8,000 additional scienceand engineering Ph.D.s needed in 1981 to 1983.There is no data on the replacement rate for in-dustrial Ph.D. scientists and engineers, but if weassume it to be equal to that of academia, or about2 percent per year, that would generate demandfor an additional 4,000 of those profefsionals be-tween 1981 and 1983. The total industrial demandfor new Ph.D. scientists and engineers wouldabout 18,000 between 1981 and 1983, while theacademic demand was 17,000.

There is evidence, however, that the industrialdemand for new doctoral scientists and engineersis not wholly supplier.: by recent Ph.D.s. Basedo:1 a comparison of N5F3 and NRC" data for the1977-83 time period, it appears that only two-thirds of the additional industrial scik.nce and engi-neering Ph.D.s reported between 1977 and 1983came from the pool of recent graduates. The restwere undoubtedly experienced Ph.D.s comingfrom academia and the government, and im-migrants.

It is difficult to predfri the future industrial mar-ket for science and engineering Ph.D.s with anydegree of certainty in the absence of an under-standing of the causes of the dramatic growth ofthe past decade. It is clear, however, that if therecent growth rate were to continue for anotherdecade it would generate substantial employment

"Ibid., tables 3 and 4, pp. x-xiii"Science and Engineering Personnel: A National Overview,op.

cit., table b-12a, p. 113."National Research Corncil, Science, Engineering, and Human-

ities Doctorates in the United States: 1983 Profile, op. cit., table2-8, p. 37.

72

Table 3.3.- Selected Characteristics of Doctoral Scientists in Business and Industry

Characteristics

1973 1983

Percentof total

employed

MedianannualsalaryNumber Percent

Percentof total

employed

Medianannualsalary Number Percent

Total .... 53,403 100.0 24.2 23,300 113,463 100 0 30 7 47,000Field:Scientists .... . 35,631 65 7 19 3 23,300 76,963 69 6 25 7 45,500Physical scientists . 19,665 36 8 40.5 22,700 28,748 25 3 44 9 45,900

Chemists . .. . ... 15,759 29 5 51 2 22,500 22,525 19 9 54 5 45,600Physicists/astronomers 3,906 7 3 22 0 23.600 6,223 5 5 27 4 48,300

Mathematical scientists 864 1.6 7 1 24,100 2,027 1 8 12 4 42,700Mathematicians . .. 657 1.2 6 2 23,900 1,512 1 3 11 1 43,600Statisticians 207 0 4 14 1 25,100 515 0 5 18 5 40,000

Computer/informationspecialists 1,007 1 9 37 1 22,300 6,819 6 0 56 1 42,700

Environmental scientists 2.204 4.1 21 4 22,600 5,154 4 5 31.3 48,500Earth sciertists ... . .. 2,085 3 9 24 4 22,500 4,596 4 1 36 7 49,200Oceanographers 94 0 2 8.3 (al 217 0 2 12 5 (a)

Atmospheric scientists 25 (b) 3 9 (a) 341 0.3 15 5 48,600Life scientists 7,147 13 4 12 6 23,300 16,444 14 5 17 7 43,700

Biological scientists .. . . 3,309 6.2 9 0 22,800 7,730 6 8 14 0 41,030Agricultural scientists . . 1,718 3 2 18.7 21,600 3,583 3 2 24 6 40,100Medical scientists .. . .. 2,120 4 0 19.9 26,300 5,131 4 5 22 2 50,700Psychologists .... ... . 3,081 5 8 12 4 30,000 13,020 11 5 27 9 48,000

Social scientists .. . . 1,663 3 1 5.7 27,200 6,751 5 9 11 4 45,400Economists .... 1,031 1 9 10 7 29,200 2,779 2 4 16 4 52 100Sociologists/

anthropologists .. 125 0 2 1 9 (a) 801 0 7 6 6 36,300Other social scientists 507 0 9 3.8 25.200 3,171 2 8 10 5 35 600

Engineers ...... .. .. . . 17,772 33 3 49 7 23,200 34,500 30 4 56 1 49,900Aeronautical/astronautical

engineers . . . . . . 639 1 2 38.3 25,300 1.928 1 7 52 3 47,700Chemical engineers . . . . 3,246 6 1 72 8 22.700 4,788 4 2 68 5 52.400Ci.il engineers . 883 1 7 28 5 21,500 1,895 1 7 35 6 47,600Electrical /electronic

engineers .. . . 3,424 6 4 49 5 23,600 7,615 6 7 60 0 51,200Mechanical engineers I 366 2 6 41 9 22,300 2,596 2 3 45 9 48,100Nuclear engineers . )66 1 2 52 7 23,100 1,380 1 2 59 3 46,200Other engineers 7,548 14 1 50 4 23,600 14,298 12 6 57 5 48.700

Sex:Men .. . 52,040 97 4 25 6 23,300 103,2"/ 2 91 0 32 2 47,900Women 1,363 2 6 8 1 20 200 10,191 9 0 20 9 38,900Race:White . 48,926 91 6 24 2 23,600 97,673 86 1 29 7 48,000Black ... 298 0 6 14 6 23.000 675 0 1.. 13 6 43,900Asian/Pacific Islander 3,099 5 8 30 0 20,800 13,431 11 8 45 2 44.800American Indian/Alaskan

native . 5 lb) 5 '°' 69 0 1 16 5 la)

Other . . . 31 0 1 15 8 1,31 76 0 1 37 6 {a)

No report . 1,044 2 0 19 3 23 000 1,539 1 4 27 7 39,800Ethnicity:Hispanic 264 0 5 16 8 21 200 1,4"1 1 3 27 9 47,230Nonhispanic 16,859 31 6 20 2 24 200 97,974 es 3 30 9 47.200No report 36,280 67 9 26 8 22 ,..- 0 13,992 12 3 29 6 45 200Ago:Under 30 1,999 3 7 20 6 17 800 2,037 1 8 31 5 36.40030-34 12,267 23 0 24 7 20,100 16.432 14 5 34 1 38,70035.39 . 10,435 19 5 24 8 22 500 25,491 22 5 33 6 44,10040.44 8,520 16 0 24 1 24,900 26,059 23 C 32 b 50,11)045.49 . 7 098 13 3 23 7 26 800 15,900 14 0 30 7 50 90050.54 . 6,288 11 8 26 1 27 800 10,65G 9 4 26 8 53 10055-59 . . 3,874 i 3 24 9 27 600 7,451 6 6 23 4 55 800

73

Table 3-3.-Selected Characteristics of Doctoral Scientists in Business and Industry-Continued

Characteristics

1973 1983

Number Percent

Percentof total

employed

Medianannualsalary Number Percent

Percentof total

employed

Medianannualsalary

60-64 1,957 3.7 21.7 27,900 6,153 5.4 27.3 53,80065 and over 932 1 7 19.1 25,800 3,218 2.8 26.5

6i300No report 33 01 33.0 la) 66 0.1 26.1

Typo of employmentScience /engineering. 49,398 92.5 23.9 23,100 98,407 86.7 30.1 46,700Other/unknown field 4,005 7.5 28.5 25,700 15,056 13.3 35.9 50,400

Primary work activity:Research and development 23,758 44.5 33.2 21,300 46,525 41.0 37.3 43,600

Basic research 3,525 6.6 10.3 21,600 6,-;31 5.9 11.8 42,100Applied research 13,212 24.7 46 0 21,300 23,463 20.7 49.5 44,100Development/design 7,021 13.1 82.6 21,100 16,331 1/ 4 80.5 43,600

Management/administration 19,840 37.2 43.0 27,000 25.528 22.5 41.3 55,900Of R&D 14,242 26.7 54.3 26,300 20,066 17.7 63.9 55,600Of other 5,598 10.5 28.1 29,400 5,462 4.8 18.0 60,C)0

Teaching 191 0.4 0.2 (a) 1,301 1.1 1.2 41,;00Consulting 2,781 5.2 68.6 25,100 10,673 9.4 83.7 48,300Sales/professional services 3,001 5.6 37.2 25,400 15,581 13.7 52.3 48,400Other 2,988 5.6 43.0 21,700 12,303 10.8 50.9 44,500No report 844 1 6 22.9 23,200 1,552 1.4 20.0 50,000allo median computed for groups with fewer than 20 individualsbLess than 005 percentNOTE Percents may not add to 100 because of rounding

SOURCE Characteristics of Doctoral Scientists and Engineers in Me United Stet,' 1983 (Washington, DC

opportunities for doctoral scientists and engineers.Assuming that the 7.8-percent growth rate experi-enced in the 1970s continues through the 1980s,and that new science and engineering Ph.D.s con-tinue to account for two-thirds of that growth,as they did in 1977 to 1983, there would be about7,000 new doctoral scientists and engineers hiredby industry each year between 1983 and 1988, and10,000 hired each year between 1988 and 1993.Both of those figures exceed the projected aca-demic demand for those years by a considerableamount. When added to the projected academicdemand for the two time periods, they lead to thecombined demand for new science and engineer-ing Ph.D.s in the two sectors between 1983 and1993 shown in figure 3-15. As can be seen, thetotals for 1983 to 1993 are quite close to that ofthe 1981-83 time period. Thus, the decline in aca-demic hiring over the next decade could be com-pletely compensated for by the increase in indus-trial demand.

Beyond the next decade, predictions of indus-trial demand for Ph.D. scientists and engineersbecome even more speculative. T!.,.! continuation

National Science Foundation, 1985), pp xii and XIII

of the 7.8-percent growth rate for more thananother decade seems unlikely for two reasons.First, the nu' it Ts of new Ph.D. scientists andengineers required for 8 percent per year growthfor another decade become so large they are im-possible to believe (the industrial sector alonewould require twice as many new Ph.D.s per yearas all sectors required in 1983). Second, the surgein growth in industrial science and engineeringPh.D.s is a relatively recent phenomenon. Statis-tics published by NRC39 show that the percent ofnew Ph.D.s planning employment in business andindustry declined steadily for all the science fieldsthrough the 1960s, and only began to increase in1973. "Until recently the trend has been down-ward for empkyment of new science Ph.D.s inbusiness and industry" NRC wrote in 1978. ISa change coming?" it wandered. If the trendreversed itself in 1973 it seems entirely possibleit could change again in another decade.

"National Research Council, A Century of Doeorates (Wash-ington, DC: National Academy of Sciences, 1978), o. 81.

74

20,000

15,000

10,000

5,000

0

Figure 3.15. Neer Ph.D.'' 'NI: Science and Enginedring

Mm,

198 1983-88 1988-93 1993-98

Five-year Intervals

ElEducational institutions Business and ildustrySOURCES: 011Ica of Technology Assessment.

The simplest assumption fe. the 1993-2003 timeframe is that the rate of industrial hiring of newscience and engineering Ph.D.s remains constantat its 1993 level. Using that assumption, the com-bined academic and industrial demand increasessubstantially above the 1983 le-Yel between 1993and 1998 and grows to e xl the current annualsupply of new science and engineeruig Ph.D.s be-tween 1998 and 2003.

The Americ t Institute of Physics (AIP) has re-cently completed an in-depth analysis of the long-

2000 2005 2010

term market fcr physic* ,11.D.s that tends to sup-port the analysis presented above. The AIPprojects that the total industrial, government, andacademic Jemand for new physics Ph.D.s couldexceed the projected supply by the year 2000 un-der the most likely scenarios.30 (See figure 3-16.)

"U.S. Congress, House, 'Testimony by Daniel Kleppner, Profes-sor of Phycics, Massachusetts Institute of Technology," Commit-tee on Science and Technology, Task F..-ne on Scier ce Policy (Wash-ington, Dr. U.S. Government Pantinb Office, July 9, 1985' pp.2-37.

POSTDOCTORAL APPOINTMENTSAmong new science and engineering Ph.D.s

who do not obtain a facrlty position or enter in -dustry, the principal alternative mode of employ-ment is the postdoctoral appointment. The Na-tional Research Council defirs a postdoctoral asa "temporary" appointme primary purposeof which is to provide for continued educationor experience in research usually, though not nee-rssarily. under the supervision of a senior men-

tor," NRC gives a nttber of rationales for theincreasing percentage of Ph.D.s taking postdoc-toral appointments:"

In many areas of science and engineering, espe-cially the interdisciplinary and trans iiisciplinary

"National ReseP.:ch Council, Postdoctoral Appointments and Dis-appointments (Washington, DC: National Academy Press, 1981),p. 11.

"Ibid., pp. 80-82.

UJ

75

Figure 3-16.Possible i "vela of Supply andDemand for Physicists WI.' ..ti Eact: Five Year

Period, 1931-20014,250

4,000

3,750

3,500

3,250

3,000

2,750

2,500

01981-88 1988-91 1991-98 1998-2001

Five-year periods

Sup ly - S

Best-,i

estimate \,

Demand - D

Bestestimate

SOURCE. American institute of Physics, In Testimony by Daniel Kieppner to theHouse Committee on Science and Technology, July 9,1905, p 52.37

ones, the nature of research has become increas-ingly complex, and has required young investi-gators to develop highly specialized skill Fre-quently eiese skills can be acquirea more effectivelythrough an intensive postdoctoral apprenticeshipt h a n t h r o u g h a graduate r e s e a r c h assistantship . . . .

From the perspective of the young invest4n.torthe postdoctoral nppointmet,. . . . provi' ; aunique opportunity to concentrate on a ticu-lar research problem without the burden or eitherthe teaching or the administrative responsibilitiesusually given to a faculty member . . . . As the

comnztition for research positions has intensi-fier', . . . . the opportunity -s a postdoctoral toestablish a strong record of research publicationshas become increasingly attractive to many youngscientists interested in careers in academic re-search . . . .

However, they .6 now considerable evidence,and concern, the t the postdoctoral appointmenthas become so nething of a holding pattern fornew Ph.D.s who cannot find 1.......ediate facultyor industrial positions. A recent Higher Educa-

10 -A

tion Resear:.h Institute survey of doctoral degreeholders who had taken postdoctoral appointmentsfound t the number of respondents reporting"employment not available elsewhere" as the rea-son for taking a postdoc it -reared from 5.6 per-cent of the 1960 to 1967 graduates to 36.8 per-cent of the 1970 to 1973 graduates. The numberciting "to become more employable" as a reasonjumped from 19.4 percent in the former periodto 39.0 percent in the latter.39

The number of science and engineering Ph.D.swith plans or firm commitments for postdoctoralstudy immediately following receipt of the degreehas increased dramaticey over the past quartercentury. Figure 3-17 shows the "percent of Ph.D.recipients from U.S. universities planning post-doctoral study immediately after receiving thedoctorate" between 1958 and 1982, by field." Itcan be seen that in the biosciences, chemistry,physics, and astronomy the percent reportingpostdoctoral plans has increased from approxi-mately 10 percent in 1958 to between 40 and 60percent in 1982. (It should be noted that data from1958 to 1969 are not completely c. oarable ,othose from 1969 to 1983 because of a change insurveying procedures at NRC.) In biochemi.the percentage today exceeds 70 percent, whi.e inmost other science and engineering fields it is rela-tively low (10 to 20 percent).

According to NRC, postdoctoral appointeesrepresented 3 percent of the U.S. doctoral scien-tific and engineering labor force of .?,65. 3 in1983, or 11,000 individuals." This was an increaseof 5 percent from the 10,500 reported by NRCfor 1979.42 More than 6,600 of the 1983 postdoc-toral appointees, or 6G percent, were life scien-tists. They represented about 44 percent of the15,000 Ph.D.s awarded in the life sciences in the1981-83 period. (The average postdoctoral ap-pointment lasts 3 years.) The next largest group,

"William Zumeta, "Anatomy of the Boom in Postdoctoral Ap-pointments During the 1970s: Troubling Implications for QualityScience?" Science, Technology and Human Values, vol. 9, No 2,spring 1984, pp. 23-37.

"William Zumeta, Extending the Educational Ladder: The Chang-ing Quality and Value of Postdoctoral Study %Lexington, M4 D.C.Heath k Co., 1%5), p. 7.

"National Research Council, Science, Engineering and Humani-des Doctorates in the U.S.: 1983 Profile, op. at., pp. 28-29.

"National Research Council Postdoctoral Appointments, op. titp. 13.

2,200, or 20 percent, were in the physical sciences.They represented about 22 percent of the physi-cal science Ph.D.s over the 1981-83 period. Thesocial scientists at 1,200 were the third largestgroup. Engineers and mathematicians were far be-hind at 280 and 120 postdoctorals respectively.

postdoctoral appointees with Ph.D.s from U.S.graduate schoolsthe population discussed above

-represent only half of the total postdoctoral ap-pointees in the United States. According to NSF"the to:al number of postdoctoral appointees in1983, including those with foreign Ph.D.s and firstprofessional degrees from U.S. schools, was20,800, up 13 percent from a total of 18.500 in1976." Of these postdoctoral appointees, 15,000,or 73 percent, were suppo, ..ed by the Federal Gov-ernment, with 10,000 or 50 percent on researchgrants. Since this chapter is concerned primarilywith the fate of scientists and engineers who re-ceive Ph.D.s from 'versities, the 11,000postdoctoral appoii -1:1 by NRC is themore relevant population.

In 1981 NRC completed a conprehensive studyof "Postdoctorals in Science and Itngineering inthe United States," which examined in depth therole of the postdoctoral appointment in the tran-sition from graduate school to the workplace."It found that between 1973 and 1979, 29 percentof the science and engineering Ph.D.s who enteredthe labor force each yea- (4,300 4. ut of 15,000)took postdoctoral appointnwnts. hi the biosciences,55 percent of the new Ph.D s who entered the la-bor force each year (2,000 out of 3,600) took animmediate pf sstdoctoral appointment. In physicsand chemistry, the fields with the next highestproportions of postdoctoral appointments, thecorresponding percentages were 46 and 40 percent.

NRC also found that in the biosciences, the du-ration of the postdoctoral ,ripointment had in-creased appreciably. Fifty-st ven percent of the bi-'scientists who received their degree in 1976 andreported plans for postdoctoral study to NRC heldappointments for longer than 2 years, up from34 percent among 1969 doctoral degree recipients

'A,:ademic Science/Engineering. Graduate Enrollment and Sup-port Fall 1943, N' 85-300 (Washington, DC. National ScienceFoundation, 1985), t..bla A-33, p 52.

"Ibid., table C-35, p. 138"National Research Council, Postdoctoral ,appointments, op. cit.

a

77

Figure 3.17. Percent of Ph.D. Recipients From U.S. Universities Planning Postdoctoral Study ImmediatelyAfter Receiving the Doctorate, Selected Disciplines, 1958-82

80

70

60

50

40

30

20

10

0

p

1958 60 82 84 68 68 70 72 74 76 78 80 82

Year of doctorate

NOTE: Data were reported only for even-numbred years Data fa intervening years were interJolated

SOURCE William Zurneta. Extending the Educational Ladder The Changing Quality anti Yalu* of l'oatdoc. oral Study (Lexington. MA D C Heath & Co. 1085, p. 7.

78

reporting planned postdoctoral study. Thirty-fivepercent of the 1975 Ph.D. recipients reportingplanned postdoctoral study, as opposed to12 per-cent of the 1968 recipients with similar plans heldappointments for more than 3 years." (See fig-ure 3-18.)

To determine the effect of the postdoctoral ap-pointment on future employment prospects, NRCcompared the 1979 employment situation of asample of bioscience, physics, and chemistryPh.D.s who received their degree in 1972 and tooka postdoctoral appointment immediately there-after, with the situation of a comparable sampleof Ph.L.s from the same fields and year whonever held such an appointment. It found that thegroup that had taken post loctoral appointmentswas far more likely to be employed at a majorresearch university and involved in research thanits nonpostdoctoral counterpart, but far less likelyto have received tenure. (See table. o-5, and3-6.) It also found that nonpostdoctorals earned3 to 10 percent more than their postdoctoral coun-terparts:. These last two effects are partially ex-plainea by the delay in obtaining a position caused

*Ibid., pp. 65-106.

Figure 3-18.Percent of Biosclentlsts Planning:vstdoctoral Study Who Had Held Appointments' anger Than 2 to 5 Years, by Year e Doctorate60

50

40

30

20

10

01967 13 1919 1970 1971 1972 1973 1974 1975 197e

Year of doctorateSOURCE National kesearch Council. Postdoctoral Apf)001fMnr8 and Disap-

pointments (Washington, DC National Academy Press, 1961). p 99

Table 3.4. Comparison of 1979 Employment Situations of Fiscal Year 1972 Bioscience Ph.D. RecipientsWho Took Postdoctoral is vointments Within a Year After Receipt of Their Doctorates With the

Situations of Other ilscal Year 1972 Graduates Who Have Never Held Appointments

Toot( postdoctorate withinyear after graduation Never held postdoctorate

Employment position in 1979 Number Percent Number Percent

Total 1072 bioscience Ph.D.s' 1,571 100 1,472 100Major research unlversitiesb 446 28 230 16

Tenured faculty 77 5 130 9Nontomed faculty 281 18 69 5Nonfaculty staff 88 6 31 2

Other universities and colleges 548 35 649 44Tenured faculty 116 7 410 28Nontenured faculty 369 24 195 13Nonfaculty staff 63 4 44 3

Konacadernic sectors 537 34 580 39FFRDC Laboratories 30 2 12 1

Government 171 41 243 16Business/industri 183 12 214 14Other sectors 153 10 111 8

Unemployed and seeking Job 40 2 13 1

jraditatse, not *dive In the labor force in 1979Illietolk ere TN universities whose total Ft.7. D expanditures in 1977 represented two-thirds of the total expondnures of all unnalraithIll and collegesNOTE: "%montage estimates repotted In this table are dertved from a sample Safvey and an subject to an absolute sampling error of Ifs. than 5 percentage points

SINNICE: National Neseeloh Council, Postdoctoral Appointments and Diseppointrnonta (Washington, DC National Academy Press. 1961). p 92

79

Table :&Cornparison of 1979 Employment Situations of Fiscal Year 1972 Physics Ph.D. RecipientsWho Took Postdoctoral Appointments Within a Year After Recoil): of Their Doctorates With the

Situations of Other Fiscal Year 1972 Graduates Who Have Never Held Appointments

Took postdoctorate withinyear after graduation Never held postdoctorate

Employment position in 1979 Number Percent Number PercentTotal 1972 physics Ph.D.s' 557 100 632 100Major research universities° 115 21 46 7

Tenured faculty 28 5 19 3! ;Mowed faculty 53 10 19 3Nonfaculty staff 34 6 8 1

Other universities and colleges 132 24 166 26Tenured faculty 49 9 113 18Nontenured faculty 63 11 41 6Nonfaculty staff 20 4 12 2

'lnseademie sectors 308 55 418 eaFFPDC Laboratories 91 16 82 13Government 71 13 107 17Business/Industry 128 23 188 29Other sectors 20 4 43 7

Unemployed and seeking job 2 0 2 0afasoludes graduates not active in the labor force In 1970.bInciuded are SS universities veal' Wel R&D expenditures In 1977 reprmentad two-thirds of the total expenditures of all universities ono co..moseNOTE: PereeMags estimate reported in this table are derived from a lamp* survey and are eubleCt to en absolute sampling error of ion Ilan 5 percentage points.MAWS: NeliSiftei aseercit stmt!!. Pcildectrraf Apodi-'-.snis aild C4PPOd"--4,Vo ciininnston. DC National Acaoemy Press. 1R11:, p 52

Table 34.Comparison of 1979 Employment Situations of Fiscal Year 1972 Chemistry Ph.D. ReelpientsWho Took Postdoctoral Appointments Within a Year After Receipt of Their Doctorates With the

Situations of Other Fiscal Year 1972 Graduates Who Have Never He.d Appointments

Took postdoctorate Withinyear after graduat:on Never held postdoctorate

Employment position in 1979 Number Percent Number PercentTotal 1972 chemistry Ph.D.s' 941 100 615 100Major research universities° 142 15 1C 3Tenured faculty 42 4 18 3

Nontenured faculty 64 7 0 0Nonfaculty staff 36 4 0 0

Other universities f:nd colleges. 119 13 106 27Tenured faculty 25 3 117 19Nontenured faculty 79 8 41 7Nonfaculty staff 15 2 9 1

Nonacademic sectors 680 72 416 68FFRDC Laboratories 38 4 0 0Government 78 8 86 14Business/industry 501 53 287 47Other sectors 63 7 43 7

Unemployed and seeking job 0 0 15 2'Excludeses g Antes not ICUs in the labor force n 1979Included re al universities whoa* total R&D expenditures in 1977 represented two-Owes of 1M total expenditures of all universities and college,

NOTE: Percentage estimates reported in this table are derived from a sample Survey and are subum to an absolute sampling loner of less than 5 percentage pointsSet -pp 4 for a description of the formula used to calculate approximate sampling errors

SOURCE National Research Council, Postdoctoral Appointments and Disappointments (Washington. DC National ACIlderr.y Press. 191311, p 92

so

by the time spent at the postdoctoral level. Themagnitude of the effect was so large in the bi-osciences, however, that NRC expressed concernover "whether the postdoctoral experience hasbeen advantageous to those pursuing careers inresearch." It reported "frustrations" among youngbioscientists caught in a "postdoctoral holdingpattern," as exemplified by the following quote:45

Frankly, many of us are concerned about ourfuture prospects in these times, after many yearsof training. We are becoming increasingly dis-couraged by the decline of tenure-track positionsand the increasing difficulty in obtaining grantsuppr,rt. An opinion that is often expressed is thatwe pcstdocs provide a cheap labor source for"established" investigators. Especially in recentyears many of us have been completely bypassedby the economic trends, so that we have beenunable to purchase homes, have families, etc.,while pursuing advanced training necessary to se-cure a " respectable position." For many of u3 itis becoming reasonable to ask: "Is it worth it?"

;liria , p. 93.

In summary, NRC found that "postdoctoralscontinue to play an important role in the nation'sresearch enterprise," and are "an invaluable mech-anism for strengthening and confirming the re-search potential of the young investigator." How-ever, NRC found "some serious concerns . . .

re- .ding the present and future role of postdoc-torals in the research community." These were:45

1. the lack of prestige and research independ-ence in postdoctoral appointments for themost talented young people;

2. the mismatch between the important rolethat postdoctorals play in the Nation's re-search enterprise and the lack of opportuni-ties that the; find f,r subsequent career op-portunities in research;

3. the lack of *et ognized status of postdoctoralappointments in the academic community;and

4. the underutilization of women and membersof minority groups in scientific research.

"Ibid.

SUPPLY OF AND DEMAND FOR ENGINEERING FACULTY

Unlike the rest of academia where the demandfor new faculty has been and will continue to beweak, engineering departments appear to besuff" g from a surplus of unfilled faculty slots.Th. se concerned with engineering education andthe health of American engineering feel that thelack of sufficient faculty is the most important de-terrent to increasing the quality, scope, and num-ber of engineenng programs. For several years,the engineering community has been referring tothe growing faculty shortage problem as the mostimmediately pressing problem facing engineeringeducation. Undergraduate engineering enroll-ments have more than doubled in the past dec-ade. In addition, graduate enrollments have nearlydoubled. hst9ntime, full-time engineering facultyrose 14 percent and part-time engineering facultyrose 20 percent." These trends have raised the stu-

"American Aftociation of Engineering Societies. Engineering andTechnology Enrollments, Fall 1;63 (New York: American Moeda-don of Engineering Societies Publications, 1984). Data was derivedfrom the 16th annual survey of engineering and technology enroll-ments conducted by the Engineering Manpower Commission ofMEN.

dent faculty ratio from 12:1 to 20:1, after adjust-ing fix 'acuity research commit"ients. (In 1973there were 235,000 full-time equivalent engineer-ing undergraduate and graduate students and19,400 FTE engineering faculty not invo!ved ex-clusively in research. By 1983 those numbers hadincreased to 486,400 FIE students and 24,800 FITfaculty.5°) If th' number of engineering majorsstays constant of slot. is in its increase, the allegedshortage will eventually abate. Engineering enroll-ment was actually down 2.8 percent in 198' from1983.5'

"Academic ;. Sdence/Engineerirs: S:ientists and Engineers, NSF84-309 (Washington. DC: National Science Foundation, 19113), ta-ble 8-1, p. 6 for faculty. All figures multiplied by 0.78 to eliminateFTEs devoted purely to research (same source, table A1A. p. 70):U.S. Congress. House, 'Testimony of David R. Reyes-Guerra, P.E.,Executive Director, Accreditation Board for Engineering and Tech-nology," Commiltee on Science and Technology (Washington, DC:U.S. Government Printing Offir, July 24, 1985), chart D for en-rollments.

"U.S. Congress, Howe. 'Testimony of David R. Reyes-Guerra,"op. tit., chart D.

t.

Many in the engineering community argue thatthe increase in the student to faculty ratio has sig-nificant:y reduced the quality of engineering edu-cation and may be forcing faculty to pursuenonacademic positions. Surveys conducted by theEngineering Manpower Commission for the Engi-neering College Facult v Project reported thatfaculty shortages had resulted it. higher teachingloads and some curtailment of course offerings."It is important to note that there is no consensusthat the rising student to faculty ratio has had amajor impact on the quality of engineering B.Srecipients. RoPert Armstrong of DuPont has pub-licly stated that industry does not see this as aproblem, as have others."

The extent of the shortage varies, depending onthe manner in which it is calculated. Accordingto NSF there are nearly 29,000 engineers employedfull-time at universities and colleges and an ad-ditional 8,800 employed part-ti:ne." A 1983 sur-vey of engineering faculty and graduate studentsby the American Society for Engineering Educa-tion (ASEE) reported that 8.5 percent of the au-thorized full-time faculty positions were unfilledin the fall of 1983 as compared to 7.9 percent in1982." However, Edward Lear, executive direc-tor of the ASEE, estimates that the shortage is ac-tually about 20 to 25 percent. He claims that torestore the ratio of faculty to students that existedin the late 1960s would requir:. an additional 6,000faculty in the engineering colleges." Lear claimsthat the true shortage is understated because ad-ministrators will not authorize positions, even ifthey ate needed, if there is no prospect of theirbeing filled.

On the other hand, Sue Berryman, a socialscientist at the Rand Corp., analyzed NSF dataon acad- .nic engineering and found a possibleshortage of computer science faculty but not of

"McPher<on, The State of Academic Labor M*.ets." op cit.,pp. 18-19

"National Re !arch Gemini, Labor Market Conditions for Engi-neers: Is There a Shor -.gel (Washington, DC National AcademyPress, 1984), :k 1, p 122.

"Academic Science/ .:nineering: Scientists and Engineers, op. at ,

p. 83, and table B-1, p. 6."Paul Doigan, "ASEE Survey of Engmeenng faculty and Grad-

uate Students, Fall 1983," Engineering Edu:ation, October 1984. p.50

"National Research Council, Libor Mark Conditions, op. cit .

p. 115.

81

electrical engineering faculty. She bases this fird-ing on an examination of comparative ,diaries,tenure rates, and rates of resignation among thedifferent science and engineering field:. Even ifthere is a shortage of computer science faculty,Berryman points out, the implications of such ashortie are uncertain. She notes that many ofthe ..ddemic requirements for computer iciencecan be fulfilled in courses other than computerscience."

Another finding that raises questions about theseriousness of the engineering faculty shortageproblem comes from the NRC survey of doctoraterecipients. According to that survey, the numberof engineering dcctorate recipients reporting def-inite postgraduation plans for academic employ-ment declined between 1973 and 1979 and hasonly returned to its earlier level in 1983. If engi-neering doctorate recipients are as much in de-mand by engineering departments as the engineer-ing professions claim, the number reportingdefinite academic plans should have increasedsubstantially.

The NRC data also reveals that the proportionof engineering doctorate recipients without firmplans at the time of receipt of the doctorate isabout the same as that for all other fields and hasremained constant over the past decade. In 1983,27.6 percent of e 4 engir ers reported that theywere still seeking appointments as their doctoralstudies were completed, the same percentage asin 1973, while 27.1 percent of all doctorate recip-ients made the same report. li the competition forengineering Ph.D.s between academia and indus-try were as strong as some claim, ti.e proportionof engineers without definite employment plansshould be lower than that for all other fields andshould be c'ecreasing."

Setting aside considerations of the magnitudeof the shortage problem, administrators claim thatengineering departments increasingly depend-ent on hold-over retirees who will soon leave thesystem entirely, on part-time personnel chosen

"Sue E. Berryman, The Adjustments of 'nu". and EducationalInstitutions to Technologically-Generated C .rages in Skill Require-ments." National Commission for Employmeat Policy, May 1985

"National Research Council, Summary Report 1983: DoctorateRecipients Frcm United States Universities (Washington, DC: Na-tional Academy Press, 1983), p. 19.

JO

82

more for their availability than for their exper-tise, and on foreign nationals (some of whom areti.Jught to be uncierqualifitd because of allegedlanguage and cultural problems). Universitiesclaim that the reasons they cannot recruit the bestengineers to teach are the relatively lower sala-ries of faculty versus those found in business andthe lack of state-of-the-art equipment for research.Daniel Drucker of the University of Florida claimsthat the autdamental problem is one of quality.Not enough of the top-quality engineering under-graduates are pursuing the Ph.D., and of thosewho do, not enough want to teach."

There are signs, however that the situation isimproving. The ASEE Report in 1983 showed that35.8 percent of institutions reported an increasein their abilit to recru: and retain facultyupfrom 16.5 percent in 1981." Unfilled engineeringfaculty positions at the professor level in 1982were only 2.9 percent, below the estimated aca-demic norm. The biggest problem seemed to ex-ist at the assistant professor level where 24.4 per-cent of the budgeted positions were reportedvacant. At higher levels there actually appears tobe a net flow of engineers into the universities;i.e., more engineers leaving industry for &Ade-mia than vice-versa. Electrical engineering experi-enced the greatest influx of new faculty membersfrom industry, with the gain exceeding the lossby more than 150 percent."

Increasingly, the rewards for university posi-tions are becoming more attractive. The Surveyof Deans conducted by AAESIASEE documentedthat many universities were providing differen-tial salary treatment for engineering faalty forpurposes of recruitment and retainment. Increasesawarded to engineering faculty were over andabove normal, regularly scheduled university-wide salary adjustments. Engineering faculty sal-aries have risen 7 to le percent per year recently,while nonacademic engineers have received raiseson the order of 2.8 percent per year."

"National Research Council, L. or Market .7onditiors, op. cit.,p. 113.

'Voipn, op. cit., p. 52."American Association of Engineering Societies/American So-

ciety for Enginee:ing Education, Final Report: Engineering CollegeFacuk Shortage Project (Washington, DC: American Society forEngineering Education, November 1983), p. 10.

"National Research Council, Labor Market Conditions, p. 118.

The relationship of academic salaries to nonac.demic salaries is illustrated in figure 3-19 show-ing academic and industrial salaries for engineer-ing Ph.D.s. The figure shows that only supervisoryengineers receive higher salaries than their aca-demic counterparts at the full professor level, butthat junior academics do not fare nearly as wellas junior industrial engineers. Figure 3-19 displaysmedian academic salaries for a 9-mcnth penod,augmented by a 2-month summer grant, whichis received by most faculty. These data may hedeceiving, as they do not include cmpensationreceived by faculty for consulting. Robert Weather-all of MIT reports that electrical engineering grad-uates who received Ph.D.s in 1967 and joinedfaculties were receiving more financial compen-sation from all sources, including consulting, thantheir counterparts in industry. Weatherall con-cludes that graduate students are disturbed moreby a lack of equipment and resources than by sal-ary differentials."

A survey conducted by the College and Univer-sity Personnel Association found that engineer-

Ibid., pp. 118-119.

Figure 3.19. Comparison of Academic-industryEngineering Ph.D. Salaries (All Professorial Salaries

Adjusted to 11-Month Basis)

64,01"1

80,000

56,000

52,000

48,000

r 44,000

Z 40,000

I moo1

32,000

MOO

24,000

20,000

16,0000 3 6 8 12 15 18 21 24 27 30 33

andYears since baccalaureate degree over

SOURCE Engifteenn9 Manpower Commission, ARES. 1987

83

ing faculty earned higher salaries in public insti-tutions than in private institutions during the1983-84 school term." Yet, according to the ASEEsurvey reported above, there were more unfilledvacancies in the public schools than in privateschools. This finding also calls intc question theclaim that the salary differential is the principaldisincentive to a career in academia.

Many in the engineering c immunity believethat the number of doctorate awards must be in-creased in order to reduce the impacts of short-ages of engineering faculty. Fortunately, the num-ber of doctorates awarded anrually in engineeringhas been rising steadily since 1980.* Engineeringdoctorate degrees inaeased by 4.8 percent in1984," when all other disciplines, except for thephysical sciences, experienced dedines." Full-timeenrollment of U.S. graduate students in engineer-ing has risen for the past 3 years."

Some question the extent to mini& undergradu-ate engineering education is necessarily dependenton doctoral level faculty to accomplish its mis-sion. The Panel on Engineering Graduate Educa-tion and Research of the National Research Coun-cil Committee on the Education and Utilizationof the Engineer concluded in their 1985 report thatschools that emphasize undergraduate educationcan safely utilize faculty without the doctorate.However, the "faculty in research universities

"Scientific Manpower Commiosion, Manpower Comments, vol.21, No. 3, April 1%4, p. 16.

The percentage of foreign nationals enrolled in doctoral engi-neering programs has remained constant at approximately 41 per-cent during this time period.

"Patrick Sheridan, "Entneering Enrollments, Fall 1983," Engi-neering Education, October 1954, p. 46.

"National Research Council, Doctorate Recipients, nku, op. at.,see table A and figure 2.

"Doigan, op. dt., n. 54.

should, in the overwhelming majority, have doc-tor's degrees.""

In conclusion, there appears to some disagree-ment about the actual need for academic engi-neers. Documented shortages are much lower thanshortages determined on the basis of some idealquality of education. In addition, the shortagesare more apparent at some faculty levels, particu-larly entry levels, and in particular disciplines,such as computer science. These shortages haveoccurred more because of skyrocketing under-graduate enrollments in engineering than becauseof the decline in Lite number of engineers seekingPh.D.s and subsequent academic appointments.If universities and colleges continue to allow highenrollments in engineering in response to per-ceived market signals, then faculty shortages willcontinue, particularly if academe continues to relyonly on field-specific, U.S. educated doctorate re-cipients as their primary source for academicfaculty. S one academic institutions have alreadybegun to cap their undergraduate engineeringenrollments L an attempt to relieve the facultysivrtage.

In addition, the literature suggests that onemeans of drawing more doctorate recipients intoacademe is through improvement of research fa-cilities and equipment. The salary differential doesnot appear to be as important a disincentive assome would suggest; many young Ph.D.s preferto conduct their research in industry because thefacilities and equipment there are at the leadingedge.

"National Research Council, Enguteenng Graduate Education andResearch (Washington, Dr National Academy Press, 1985), p. 103.

U.S. EDUCATION AND UTILIZATION OFFOREICN SCIENTISTS ANC ENGINEERS

Both as students and as faculty me- .ers, for-eign nationals are a significant portion of the aca-demic population, particularly in engineering,computer science, and the physical sciences. For-einm students comprised 6.2 percent of the under-gisduate enrollment in engineering, and 35 per-cent of the graduate enrollment. Their overall

participation in all of higher education enrollmentis less than 3 percent, de-nonstrating their dis-pmportionatc participation in engineering educa-tion." One-fourth of all graduate students who

"Participation of Foreign Citizens in U.S. Science and Engineer-ing (Washington. DC: National Science Foundation, January 1985).

2

84

are foreign nationals are enrolled in engineeringprograms.

There has been a steady increase in enrollmentof foreign students since 1964, when total foreignenrollment in U.S. institutions of higher educa-tion was 1.5 percent." (See figure 3-20.) A Re-port of the Institute of International Educationconcluded that foreign students enroll proportion-ately more often in engineering and the medicalsciences, and less often in the humanities and so-cial sciences This can be attributed to the unavail-ability in developing countries of costly trainingfacilities necessary for engineering and medicalscience education." Since 1979, the regional ori-gin of engineering students has changed dramat-ically. The proportion of South and East Asiansnearly doubled between 1979 and 1984, while theproportion of students from the Middle East de-dined by 17 percent. Data from 1984 show thatSouth and East Asians constitute nearly 60 per-cent of foreign graduate engineering students, and32 percent of foreign undergraduate engineeringstudents. Middle Easterners comprise 18 percentof the foreign graduate engineering populationand 38 percent of the foreign undergraduate engi-

"Ibid., p. 7."Elinor G. Barber (ed.), Foreign Student Flows, HE Research Re-

port No. 7 (New York: Institute of International Education, 1955),pp. 1042

I

Fig, s 3-20.-Total Foreign Enrollment In U.S.Institutions of Higher Education

400

XO

300

100

0

ii

r

ir

4

1154 IMO 1554 1510 1575 1550 1552

Foreign enrollmentse a percent oftotal enrollnrint

1.5 1.7 1.5 2.5 27

NOTE Inductss students at 6N levelsSOURCE Institute of International Education and National Center for E million

Statistics

neering population." Table 3-7 shows world re-gion of origin of foreign students in the sciencesand er- 'neering undergraduate end graduatelevels combined.

Data for 1983 from AAES shc'w that at thePh.D. level, 41.5 percent of total full-time engi-

ffMarianthi Zikopoulos and Elinor G. Barber (eds.), Profiles:Detailed Analyses of the Forman Student Population, 1983/1984(New York: histlute c5 International Education, 1955), pp. 3439.

Table 3-7.-Origin of Foreign Students Within Fields of Study, 10344

Worli region

Tots:

Number ofstudentsreportedAfrica Europe

LatinAmerica

MiddleEast

NorthAmerica

South anc.:Oceania East Asia

Agriculture 28.8 5.5 27.9 8.7 4 1 1.0 26.3 100.1 4,993Business and

management 18.1 8.8 17.8 12.3 4.2 1.3 39.7 37,939Education 16.5 9.7 17.1 13.2 10.8 4.1 28.7 7,944Engineering 7.0 6.5 12.2 30.1 1.8 3.2 42.2 43,281Fine and applied arts 15.8 8.5 21.2 18.5 4.8 3.5 34.5 4,185Health sciences 18.3 9.9 18.8 1 ..3 98 2.3 28.4 8.814Humanities 9.8 21.4 14.4 6.8 9.7 2.5 35.8 3,490MathIcomputer science 6.4 7.1 13.0 17.2 1.5 0 7 54.1 18,993Physiccifilfit sciences 10.8 11.4 12.8 14.3 4.3 0.9 45.7 15,888Social sciences 8.8 18.4 18.2 10.7 11.0 2.1 34.8 7.739Other 14.4 12.9 17.8 11.1 7.5 1.8 34.7 7,400:Atensive English

language 2.9 5.4 22.5 26.5 0.5 0.0 2.0 3,298Undeclared 10.3 18.1 19.1 12.8 9.3 1.3 31.2 8,108All fields 11.4 9.6 15.8 17.4 4.8 1.2 39.9 177,050SOURCE National Wanes Foundation.

r" 0s I ut

i

85

neering can aidates were foreign nationals, upslightly from 41 percent in 1982.* The NationalResearch Council Survey of Doctoral Recipients(1983) shows that the percentage of engineeringdoctoral recipients listing U.S. citizenship has de-dined from 82 percent in 1965 to just under 48percent in 1983." In comparison, in 1983, non-U S citizens received more than a third of the doc-toral degrees awarded in mathematics, 28 percentof physics and astronomy doctorates, 21 percentof the chemistry doctorates, and 12 percent of thedoctorates in the biological and health sciences(down from 14 percent in 1975).

In the doctoral population, available informa-tion from various sources indicate that between20 and 50 percent of foreign doctoral students inthe sciences are principally supported thrAigh re-search or teaching assisantships. In engineering,the proport: on is nearly 60 percent." It is uncer-tain whethe r this deprives qualified U.S. studentsof research trap ing and financial support. Ana-lysts Lewis Solmon and Ruth Beddow found that,"foreign students probably pay more than theireducation costs, and they might be paying sub-stantially more."" Sohnon and Beddow did notinclude U.S. -esearch assistantship or fellowshipsupport in their analysis, however.

Proponents of U.S. education assistance to for-eign nationals justify it with the following reasons;it is a means of helping persons from less fortu-nate nations to obtain the benefits of a U.S. edu-cation; the political advantage of having futureleaders of other countries educated in the UnitedStates, rather than elsewhere will pay off in thelong run; it is to the cultural and sociologicaladvantage of U.S. students to be exposed to per-sons from other parts of the world; and the UnitedStates can utilize the skills and training of thoseindividuals who stay on in this country after their

*In base numbers, in 1982 there were 6,741 foreign nationalsanions the 16,442 engineering doctorate enrollees; in 1983, 7,687of the total 18,540 doctoral candidates were foreign nationals.

"National Research Coundl, Summary Report 1983: DoctorateRed,riente From United States Universities (Washington, DC: Na-tional Academy Press, 1983), see table D.

"Betty Vetter, "U.S. Filication and Utilization of Foreign Scien-tists and Engineers," contram or report prepared for the U.S Con-pew Office of Technology Assessment, 1985.

"Lewis C. Solmon and Ruth Beddow, "Rows, Costs and Bene-fit; of Foreign Students hi the United Stites: Do We Have a Prob-leer in Elea Barber, op. dt., pp. 121-122.

educational training has ended. Opponents areconcerned that the training and employment offoreign nationals in the United States constitutesa "brain-drain" from the home country if they re-main K. and could pose a threat to our nationalsecurity, if they return hc:ne. Restricting accessto "sensitive" research at U.S. universities mayprovide additional problems to those research in-Aihitions with large numbers of foreign studentsand faculty members.

There is disagreement between various groupsin this country as to whether indHduals admittedto the United States as students who complete agraduate degree here should be allowed to remainafter completion of the degree (and perhaps ashort period of further training). Increasingproportions of those earning doctorate degrees (56percent in 1983) are remaining in the country af-ter graduation." Many employers, both in indus-try and in academic institutions, say that they can-not fill certain research and faculty positions withU.S. citizens, and avail; data verify this.77 In1984, major changes in itionig,mion rules wereproposed in legislation passed separately by bothhouses but not signed into law, both of which in-cluded an exception to the general rule that allforeign nationals in this country on a student visabe required to return to their native countries af-ter graduation for 2 years before attempting toreenter the United States on a permanent visa. Theexceptions would have allowed some scientistsand engineers, as well as some other specialists,to accept employment and remain here. Univer-sities and some industries felt sufficiently depend-ent on the foreign national population with U.S.graduate degrees that they lobbied Congress toinclude the exception allowing scientists and engi-neers to stay and accept employment in the UnitedStates."

Others are concerned that the increasing num-ber of foreign engineers earning degrees in thiscountry are taking jobs at low pay in order to re-main here, thus driving down salaries and redur.-ing opportunities for U.S. members of the engi-

"Participatios of to.-.74.tn Citir ms, op. dt., p. v."M; -hail G. 'Inn, "Foreign i-itional Scientists and Engineers in

the U . Labor Fori.., 1;77-1982," a report to the National SdenceFoundation, March 1985, draft.

"Vetter, op. dt., p. 16.

86

neering profession. There is no statistical evidencehat foreign engineer are paid less than compara-

bly trained and expt.rienced U.S. engineers."

Foreign nationals are employed proportionatelymore often in educational institutions, specificallyhigher education, than U.S. citizens. This may beexplained by the fact that scientists and engineerswho are foreign nationals tend to have higherlevels of educational attainment than U.S. scien-tists and engineers, as shown in figure 3-21. Theproportion of foreign nationals engaged in re-search, development, and design is about the sameas for U.S. scientists and engineers. The propor-tion of foreign nationals in management positionsin R&D firms, however, is less than their U.S.counterparts. This may be a reflection of the factthat foreign nationals in the U.S. science and engi-neering labor force tend to be younger than theaverage U.S. scientist or engineer. Managementpositions traditionally go to senicz employees.°0

Many foreign nationals choose to enter the aca-demic market after graduation. Because there isa shortage of faculty in engineering and in com-puter science and because an increasing propor-tion of U.S. doctoral graduates in these fields areforeign citizens, U.S. faculties include a large andincreasing proportion of foreign nationals and for-eign born but U.S. educated citizens. The currentshortages of faculty in engineering schools wouldbe far higher had they not been able to employforeign engineers with U.S. Ph.D.s. A 1982 sur-vey found that 18 percent of all full-time engineer-ing faculty members earned their B.S. degreesfrom an institution outside the United States. Thehighest percentage of foreign born faculty is incomputer science/engineering-20.7 percent.More than one-fourth (26 percent) of all assistantprofessors in engineering in 1982 earned their B.S.degrees outside the United States.'

Several authors have made reference to the",rowing problem" of foreign nationals on engi-neering faculties. They contend that the draw-backs of an increasingly foreign born faculty are

"Vetter, op. cit., p. 17."Finn, op. cit., p. 12."Vetter, op. cit.

rt i j

Figure 3.21.Educational Attainment of Scientistsand Engineers, by immigrant Status, 1982

Other

U.S. natives

Other

Foreign natiooafe

Other

Naturalized U.S. cit.mnsSOURCE ORAU, based on 1982 Postcensal survey (Finn)

the limitations in their communications skills andthe implicat;ons for education. Some studies havefound that a number of problems for women aregreatly exacerbated when they deal with foreignfaculty and foreign students." The classroom cli-mate for women who must deal with foreign malefaculty and graduate students is particularly dif-ficult when such faculty and students come fromcountries where women, by statute or custom,have a very restricted role. These difficulties arefurther emphasized by the fact that 92.5 percentof all foreign students in engineering are male.

In conclusion, there is no apparent national pol-icy in regard to either the education or the utili-zation of foreign students and graduates, againstwhich either current or proposed regulations for

"Scientific Manpower Commission, The International Flow ofScientific and Technical Talent: Data, Policies and Issues, proceed-ings of a joint meeting of the Scientific Manpower Commission andthe Engineering Manpower Commission on May 7, 1985, in prep-aration for publication, 1985.

87

the temporary or F Tmanent entry of foreignscience and engineering students or foreign scien-tists and engineers might be tested. Foreign stu-dents make up a significant iraction of graduateenrollments in U.S. universities, where they serveas research assistants, and to a lesser degree. asteaching assistants. As these students obtain ad-vanced degrees, many seek to remain in the UnitedStates and to enter the U.S. labor force, whichappears to be problematic only in academe, wherecommunication skills are essential. There is nodoubt that foreign born scientists have enrichedU.S. accomplishments and achievements in scienceand technology. As these individuals increasinglyenter American industry and academe, followingan education in American colleges and universi-ties, a new set of policy issues arise. Issues of in-teiest to policymakers are the purposes of edu-cating students from other nations, whether theeducation of Amer-can students suffers from theparticipation of foreign nationals, and the cost,if any, to the taxpayers of educating foreignstudents.

Chapter 4

The Industrial Marketfor Engineers

PageShortages: Present and Future . . 92Supply and Demand Pro ections . 95

Supply Projections ......... ... . 96Supply-Demand Gaps 97Defense Needs 97

Adjustments to Supply-Demand Gaps .. ...... ... . 99Supply Adjustments .100Occupational Mobility 101Demand Adjustments . ..102increased Search and Recruitment 102Internal Training and Retraining of Engineers 102Riartangement of Jobs 102

flit cottrequences of Adjustment 103in Shortagea of Engineers 104

of Shortages 105and Flexibility of Engineers and Related Professionals 106Changes 107

108,

,Lict of FiguresFigure No Page4-1. Percent of All Offers and Percent of All Bachelor's Graduates . ..... . . 914-2. Employed Engineering Personnel, 1960-8i 924-3. Engineering Freshman Enrollments, B.S., M.S., and Ph.D. Degrees ... . 934-4. B.S. Engineering Degrees, 1968-83 .... . ... . 944-5. Supply and Demand for L..agineers 1978-90 . .. 98

Chapter 4

The Industrial Market for Engineers

As shown in chapter 2, most science bachelor'sdegree recipients either go on to graduate schoolthe Ph.D. being the true entry-level degree fora research careeror obtain employment in anonscience field (see chapter 2, table 2-2). Morethan three-quarters of all engineering B.S. recip-ients, by contrast, obtain immediate employmentas engineers in industry. In addition, more than90 percent of all the job offers to all science andengineering bachelor's graduates in 1984 were inengineering, engineering technology, or computerscience (see figure 4-1).2 Therefore, when consid-ering the utilization of science and engineeringbachelor's degree holders, it is appropriate to fo-cus on the industrial market for engineers.

'Opportunities in Science and Engineering: A Chartbook Pres-entation (Washington, DC: Scientific Manpower Commission, No-vember 1984), p. 32.

According to the Bureau of Labor Statistics(BLS), the number of employed engineers in theUnited States doubled between 1960 and 1982, in-creasing from 800,000 to more than 1,600,000 (seefigure 4-2).2 During that same period, 1960-82,the U.S. gross national product (GNP) increasedby 100 percent in constant dollars and the nationalresearch and development (R&D) budget grew by103 percent in constant dollars. Thus, the growthin demand for engineers appears to correlate verywell with growth in GNP and growth in total na-tional expenditures for R&D.

'Engineering Education and Practice in the United States (Wash-ington, DC: National Research Council, 1985), p. 86. It should benoted that there is a serious discrepancy bewren BLS and NSF dataon employed engineers, the latter showing 1,847,000 engineers em-plcyed in 1982, Science Technology Resources, NSF 85-305 (Wash-ington, DC: National Science Foundation, January 1985), p. 545.

Figure 4-1.Percent of All Offers and Percent of All Bachelor's GraduatesMen Women

(percent) (percent)70 50 50 40 30 20 10 0 0 10 20 30 40 50 60 70

Engineering/engineeringtechnology

Business/management

Computer scl.!mathematics

Humanities/social sm.

Healthprofessions

lifesciences

Physicalsciences

SOURCE. Opportunities in Science and Engineering (Washington, DC Scientific Manpower Commission. 19114). p 32

alaJ91

92

Figure 4-2.Employed Engineering Personnel,1980-82

1,000

1,400

1,300

1,000

SOO

000

400

200

0

e

1000 1970 197419761978190019b2

Year

9099ce: Engineering Education and Practice ,n the United States (Wash-avian, IX' National Rosaarch Caused. 1985). P 88

The number of first year enrollments in engi-neering schoc and the number of engineeringbachelor's degree recipients has fluctuated quitewidely since World War II, with peaks andtroughs apparently produced by transitory polit-ical and social events and trends (see figure 4-3).3

'Engineering Educaticn and Practice in the United States, op. cit.,p. 52.

However, beneath the fluctuations, there appearsto be an overall trend of an increase of slightlyless than 3 percent per year in the number of engi-neering B.S.s awarded each year. The recent surgein engineering B.S.s awarded in 1980 to 1983 ap-pears to be something of an aberration, that couldbe followed by a decline in the late 1980s, sinceall previous peaks have been followed by troughs.

The growth in engineering B.S.s has been some-what uneven across fields, with electrical and me-chanical engineers showing the greatest increasesover the past decade, as can be seen from figure4-4.4 The number of aerospace, chemical, indus-trial, and mining and petroleum engineering B.S.shave also increased dramatically, more than dou-bling for each field since 1976. Only civil engi-neering failed to show a dramatic increase in the1976-83 period, perhaps because it was the onlyfield not to suffer a decline in the early 1970s.

'Robert C. Dauffenkach and Jack Fionto, "The Engineering De-gree Conferral Process. Analysis Monitoring and Projections," re-port for the Engineering Manpower Commission, November 1984,p. 13.

SHORTAGES: PRESENT AND FUTURE

The principal industrial employers of engineersare companies that make transportation equip-ment (11 percent); communication equipment (6percent); electronic components (3 percent); andoffice, computing, and accounting machinery (5percent). Other major employers include engineer-ing companies (10 percent) and the government(12 percent). Many engineering employers havereported shortages of engineers, especially in theelectrical and computer specialties (CS). The Na-tional Science Foundation (NSF) annually surveysabout 300 major industrial employers of engi-neers. The table below reports the percentage ofsurveyed companies reporting a "shortage" of ei-ther electrical or computer engineers.' (NSF de-fines a slt".rtage as a situation where an employerhas "more vacancies than qualified job applicantsin a specific field.") Note the dramatic decline in

"Shortages Increase for Engineering Personnel in Industry,"Science Resource Studies Highlight (Washington, DC: NationalWears Foundation, Mar. 29, 1985).

the number of firms reporting shortages between1981 and 1983.

Electrical engineers Computer engineers1981 . . 57 percent 51 percent1982. . . . . 31 percent 21 percent1983..... . .. 7 percent 6 percent1984.. 19 percent 15 percent

Other surveys report similar results. The 1983American Electronics Association (AEA) surveyof 815 firms employing engineers found that 32percent reported a shortage of electrical and CSengineers.° A similar study conducted amongMassachusetts electrical engineer (EE) employersin 1983' found 25 percent reporting shortages of"entry-level" electrical engineers, while 65 percentreported shortages of experienced EEs.

'Technical Employment Projections, 1953-87 (Palo Alto, CA:American Electronics Association, July 1903).

'Paul Harrington, et al, "Employment Developments in the Elec-trical Engineering Labor Market of Massachusetts and the U.S.,"Northeastern University, 1983.

100

Figure 4.3. Engineering Freshman Enrollments, B.S., MS., and Ph.D. Degrees

120,000

105,000

90,000

ip 75,000

80,000

I45,000rn

30,000

15,000

1945 1950 1955 1980 1965

Year

1970 1975 1900 1995

1 Returning World War II veterans2 Diminishing veteran pool and expected surplus of engineers3 Korean Wer and increasing R&D expenditures4 Returning Korean War veterans5 Aerospace program cutbacks and economic recession6 Vietnam War and treater space expenditures7 Increased ;:tudent interest in social-program careers8 Adverse student attitudes toward engineering, decreased space and oefente

expenditures, and lowered college attendance9 Improved engineering job market, positive student attitudes toward engineerinc, and entry

of nontraditional students (women, minorities, and foreign nationals)10 Diminishing 18year-old pool

A ASEE Evaluation Report recommends greater stress on math/science and quality graduateeduce ion

SOURCE Engineering Education and Practice in the United States (Washington, DC ri atonal Research Council, 1985). p 52

1A. I

94

I.8 10,000

1 8,000z

20,000

WOO

16,000

14,000

12,000

6,000

4,000

Z000

0

Figure 4-4.B.S. Engineering Degrees, 1966-63

1

A

4

1970 1975

Year

1980

SOURCE: Robot Douttienbacti and Jack Pronto, "Tn. Engineering Dogma Conform Proem Analysis, Monitoring and Protections," raport to Engineering ManpowerCommission, 1964, p. 13.

The key question for policy purposes is whethershortages of engineers will be a problem in thefuture. This question is especially difficult to an-swer when it is not clear how much of a problemshortages cause at the present time. Shortagereports do not describe the adjustments that em-ployers and potential employees make to short-age situations. Employers have many optionsavailable to them, including hiring less qualifiedpersonnel and training them; raising wages to bidaway engineers from other companies; and re-ananging jobs and tasks to better utilize engineer-

ing +alent. It is clear that such adjustments takeplace, but there is less clarity about the conse-quences. Some adjustments may be relativelypainless, while others may create as many prob-lems as the original shortages did.

Employees also have options available for deal-ing with shortage or surplus situations. They can,within limits, move to related and more profit-able fields, or seek training in those specialtieswhich are experiencing shortages. The interfieldmobility of scientists and engineers, and relative

1i i 4,3'r

95

responsiveness of underyaduates to market sig-nals, are two important characteristics of the U.S.scientific and technical work force.

In order to evaluate the possibility and effectsof engineer shortages over the rest of the decade,therefore, it is not enough to simply project sup-ply and demand. It is also necessary to understand

how employers and employees are likely to ad-just to shortage and surplus situations. The prin-cipal purpose of this chapter is to lay out a frame-work for understanding these adjustments, andto summarize the (admittedly skimpy) evidenceon their effects.

SUPPLY AND DEMAND PROJECTIONS

Supply and demand projections for engineersare the raw material for evaluating the extent offuture engineering "shortages." In February 1984,the Office of Scientific and Engineering Person-nel of the National Research Council (NRC) helda symposium on Labor-Market Conditions fo-Engineers: is There a Shortage?' This symposiumbrought together the major sources of engineer-ing supply and demand projections: BLS, NSF,and AEA.

All three projections used 1982-83 as their baseyear. BLS projected engineering jobs through1995, using a sophisticated variant of the "man-power requirements" approach. First, BLS pro-jected the growth rate of the whole economy andof individual industnes. Then it calculated howmany engineers would be needed to achieve thisgrowth rate. These calculations were based on his-torical patterns of engineering employment, plusanticipated changes in these patterns in the future.

NSF used a similar methodology to predict theneed for scientific, engineering, and technical per-sonnel over the period 1982-87. Its model, how-ever, gives more attention to specifying the courseof defense spending than does BLS.

AEA took a totally different approach. They sur-veyed their members and asked them to project

'La For Market Conditions for Engineers. Is There a Shortagol(Washington, DC. National Reseaach Council, 1984)

personnel needsespecially for engineers--through1987. The individual needs were then totaled. Thismethodology has beer criticized, by W. Lee Han-sen and others, as being inherently unreliable.Most employers of engineers do not make firmprojections of their manpower requirements be-yond the next 18 months, so the responses to theAEA survey for the out-years are largely educatedguesses. Moreover, respondents to surveys of thistype tend to be overly optimistic about the rela-tive market share their company is likely toachieve, and hence tend to overestimate their per-sonnel requirements.

It must be emphasized that all three organiza-tions were projecting the number of new engineer-ing jobs. This is not the same as the demand forengineers. Employers not only have to fill newjobs, but they also have to fill vacancies causedby death, retirement, movement out of engineer-ing into management or other jobs, or return toschool. An accepted rule of thumb is that about2 percent of the engineering work force will re-tire or die each year, creating new vacancies.Another 4 percent of engineer; will transfer outof engineering into management or some otherfield. These are not small numbers. In fact, thereplacement demand in any year is considerablylarger than the demand created by new job growth.Thus, the magnitude of any "shortage" dependscrucially on how these flows of people are treated.

....

-1. 1 . %:i

96

Under the BLS moderate scenario for economicgrowth, the number of engineering jobs would in-crease by 3 percent per year, which is identicalto the rate of growth in GNP assumed in the BLSmodel, and also to the historic rate of growth inthe engineering profession. This growth rate trans-lates into 45,000 new job openings each year.However, an additional 93,500 engineering jobswill have to be filled each year because of attri-tion and transfers.

NSF comes up with quite similar projections:an annual growth rate for engineering jobs of 2.6to 4.5 percent. The low end of the range reflectsan assumption of stagnant economic growth (realGNP growth rate of 2 percent) and low defensespending, while the high end reflects strong eco-nomic growth (real GNP growth of 4 percent andhigh defense spending).

AEA projections for demand are much higher:about 50,000 new engineering jobs created eachyear within the electronics industry alone. Ac-cording to AEA, the electronics industry employsabout one-third of all engineers. Consequently,AEA projections of demand, on an economy-widebasis, are three times higher than BLS and NSFestimates. These high projections clearly resultfrom AEA's use of a questionable methodology.

Supply Projections

There are two major compments to supplynew engineering graduates and transfers fromother occupations. All three projections agree onthe likely supply of entry-level engineers.

BLS and MA project that the supply of entry-level engineers going into industry will average63,000 annually, while NSF projects 65,000 to

69,000 annually. (All three projections fall sub-stantially below the number of engineering under-graduate degrees actually awarded in 1983 and1984 which were 72A71 and 76,931 respectively!)

The key difference is how these studies handlethe transfer component of supply. BLS allows fortransfers out of engineering (into other types ofjobs or schools) but not transfers into engineer-ing. Thus, BLS makes no attempt to calculate oc-cupational mobility into engineering (though itsexistence and importance is acknowledged).

fhe NSF study, by contrast, makes two alter-native assumptions about occupational mobility.In the first scenario, occupational mobility intoengineering (and each of its subfields) is assumedto be equal to occupational mobility out of engi-neering. In the second scenario, it is assumed thatoccupational mobility into engineering takes placeat its historic rate, which is just enough to bringsupply and demand into balance. The AEA makesno assumption about the magnitude of interfieldmobility.

Supply-Demand Gaps

What kinds of gaps between supply and de-mand for engineers do these projections imply?BLS compares the supply of new graduates to thedemand for new jobs, and for replacements dueto attrition and transfers and finds that supplyonly meets 46 percent of the demand for allengine:

'Engineering Education a,,d Practice in the United States, op. cit.,p. 57; Scientific Manpower Commission, Manpower Comments,January-February 1985, p 34.

LE

However, BLS is relatively sanguine about theimplications of this gap. It points out that in 1980,the latest year with good figures, new entrants alsofilled less than one-half of total demand. (This wasalso true over the entire decade of the 1960s,according to OTA calculations.) BLS argues thatthe remainder came from:

. . . transfers from other occupations; employedpeople with previous experience or training inengineering or a related occupation; recent scienceand math graduates; immigrant engineers; olderengineers returning to the profession . . .

BLS concludes that the labor market for engineei sin the rest of the decade should be the same asit was in 1980. So, if there were sporadic short-ages of EEs in 1980, there should be sporadicshortages in the future.

NSF arrives at roughly the same conclusion, butby a different route. Assuming that occupationalmobility into and out of engineering are equal (thefirst assumption) the NSF study projects a slightexcess of demand over supply for electrical engi-neers in 1987 (2.4 to 7.9 percent vacancies), anda rough balance for the remainder of engineeringfields. The second assumption leads, by defini-tion, to a balance of supply and demand. Finally,the shortage predictions of AEA are considera-bly more severe, due to their use of a questiona-ble methodology.

William Upthegrove presented a commentaryat the NRC symposium based on a Business -Higher Education Forum study of EngineeringManpower and Education" that sheds consider-able light on the issue of supply and demand bal-ance. Figure 4-5," taken from Upthegrove's pres-entation, shows the flow of engineers into and out

"Business-Higkr Education Forum, "Engineering Manpower andEducationFoundation for Future Competitiveness," October 1982.

"Labor Market Conditions for Engineers: Is There a Shortage?op. cit., p. 64.

97

of the prcfession. It reveals that when immigrants,bachelors of engineering technology, and B.S. de-gree holders from related fields such as physicsand mathematics are taken !nto account, onlyone-third of the growth plus replacement demandmust be filled by interfield mobility and upgrad-ing of nonenginews. Upthegrove's flowchart doesnot take into account the 18,000 new engineer-ing M.S. degree holders who enter the job mar-ket each year and are available to contribute tothe supply side of the picture. With these included,the supply-demand balance would look even morefavorable.

Defense Needs

Before proceeding to look at labor market ad-justments to shortages, a key question that mustbe examined is the effect of defense budget in-creases on the demand for engineers. How sensi-tive are the projections of engineering demand tochanges in defense spending?

AEA asked its respondents to identify the por-tion of their projected future requirements thatwere based on the assumption of receipt of de-fense contracts. The purpose of this question wasto avoid the problem of double and triple count-ing, where several firms project the same hiringrequirements based on receipt of the same defensecontracts. The response to this question can beused to estimate the effect of increased defensespending. According to the responses, 16 percentof their projected requirements were based on an-ticipated defense contracts. Hence, even a dou-bling of defense spending would not have a largeeact. This conclusion is buttressed by the resultsof Harrington, et al.'s survey which reported that"most respondents did not see increasing defenseoutlays as a key cause of future shortages.""

"Harrington, et al., op. cit.

-1 I i .-,1

98

Figure 4.5. Supply and Demand for Engineers, 1978-90

Total supply = 104,700 to 113,400 per year.Total demand (growth + death and retirement + transfers out)

= 104,100 - 113,400 per year.

Entry supply Demand c.'eatedby departures

11111111111111110

11111111111111110

Total engineering andPseudo-engineeringemploymentdemand/pool

1978-1,071,000°projected to1990-1,504,000d

am. L Carey (Bureau of Lebo, Statlallak IlkooPatIonal Employment Growth Through 1990," Monthly L ghor Rashinv, August 1901. These data from "low" growth pro-jection."Averaoss determined from values of table 19 In National Center for Education Statistics, Pro/actIons of EducationStatistics to 1988-89 (Washington, DC: U.S. Govern-ment Printing Office, April 1090); am! 1900 values from NCES Bulletin 01-406, Feb. 13,1991.wOooupatIonat tholactions and Training Data, 1960 Edition, Bureau of Labor Statistics, LIS. Department

of Labor Bulletin 2152, September 1900. The 20% figure Isprobably cOnamvadve In light of the Increasing enrollments of nonpermanent foreign nationals In (moldering programsditUltestimates. This estimate Is consistent with W. P. Upthegrove's Impression that about one-half of foreign nationals who take B.S or M S degrees In U.S. engineer-_Ing colleges eventual!? stay In the United States.

"Unpublished data compiled in 1976 -79 by Daniel E. Hecker, Bureau of Labor Statisticsfarurvwy of engineering employment 1974-1970 suggested an

avenge transfer-out of 37%. More recent surveys have Indicated an even higher rats (unpublishedBureau of Labor Statistics data from Daniel E. Heckel.

SOURCE: Labor-Market Conditions for Engineers, is There a Shortage (Washington, DC National Research Council, 1984), P 84

NSF studied the impact of different levels ofdefense spending in more detail, and came toroughly the same conclusion. It looked at twodifferent levels of defense spendingthe "high"projection which assumed a 45-percent increasein real defense spending between 1982 and 1987,and a low" projection which assumed a real in-

crease of 18 percent over the same period. TheNSF study found that shifting from low to highdefense spending projections increased the totaldemand fur engineers by 85,000 in 1987. This isan increase of about 6 percent, and has the effectof turning a slight shortage into a slightly largerone.

99

Thus, increased defense spending can boostengineering demand somewhat, but not enor-mously. This should not be surprisingonly 18percent of the Nation's engineers are employedin defense-related industries," so an increase ofone-fourth to one-third in defense spending (thedifference between the high and low scenarios)should generate a 4- to 6-percent overall increase.

There exists a rule of thumb to calculate the ef-fect of defense spending on the demand for engi-neers. According to Landis," an additional $1 bil-

"The 1982 Post Censal Survey of Scientists and Engineers, NSF84-330 (Washington, DC: National Science Foundation, 1984), p. 60.

"Fred Landis 'The E^onomics of Engineering Manpown," Engi-neering Education, December 1985.

lion (1983$) in military spending will generate ademand for about 1,000 additional engineers ifthe funds are spent on procurement, and about4,000 additional engineers if they are spent onR&D. On the average, electrical engineers shouldmake up about one-fourth to one-half the addi-tional demand.

Finally, we note that although defense spend-ing has a relative small effect on overall levelsof engineering demand, it may have a very largeeffect on specialized subfields. It is impossible,however, to generalize about these localized de-mand effects. In the next section we will discusswhat is known about supply responses to subfieldrhortages.

ADJUSTMENTS TO SUPPLY-DEMAND GAPS

The essential nature of a labor market is thatit adjusts; supply and demand respond to the sur-pluses and shortages. A projected gap betweensupply and demand usually does not indicate animpending shortage; rather, it signals that somesort of adjustment has to take place. Two typesof adjustment are possible: The first are adjust-ments among potential employees, such as an in-crease in the numbers of entry-level engineers ormobility of experienced workers into shortage oc-cupations. The second are adjustments by em-ployers of engineers who can:

increase their search and recruiting efforts;rearrange jobs to utilize available skills, edu-cation, and experience more efficiently;make larger investments on internal trainingand retraining of engineers; andcut back on production or on R&D.

Under certain restrictive conditions no adjustmentis possible. Most projections of shortages assumetacitly that demand and supply are independent,so that there can be no adjustment. This assump-tion makes sense, however, only under the fol-lowing very restrictive conditions:

demand for the final product is relatively un-affected by labor costs;supply is not appreciably affected by wagechanges; and

the shortage skills are unique, in that work-ers possessing them cannot be replaced byworkers from other occupations or by newtechnology.

If arty of these assumptions are violated, thena projected supply-demand gap will Ir9 closed bymarket adjustment. These assumptions are veryrestrictivethe latter parts of this section provideevidence to support the prevalence of adjustment.

However, there are two sectors of the engineer-ing labor market in which it makes some senseto assume that supply and demand will not ad-just. The first is the defense sector. In defense-related work, the demand for the final productis usually not heavily influenced by the cost (andthus not by the associated labor costs). Moreover,defense-related work may demand very special-ized subfields that are in short supply in the shortran.

Not surprisingly, the NSF surveys of shortages"bear out this observation. About 64 percent offirms with a high proportion of their employmentin defense-related work reported shortages, com-pared to 33 percent overall. Thus, at least in theshort run, it makes sense to project supply anddemand independently for the defense sector.

""Shortages Increase for Engineering Personnel in Industry," op.cit.

The other sectcr in which a supply-demand gapactually reflects a shortage are startups. Consider,for example, newly formed companies in the com-puter industry. Since they are not yet producinga product, they are willing to accept losses in theshort run in exchange for the possibility of largereturns in the future. Instead of being expectedto immediately make a profit, they are compet-ing to get their product to market before theirrivals. For firms such as these to cut back on hir-ing when wages go up would be completely self-defeating, s nce it would redu:e the possibility ofachieving pi ofitability at any time in the future.Thus, demand for engineers, in this case, wouldbe independent (within limits) of the wage level.Similarly, if the firm is using cutting-edge teell-nology, the engineers skilled in this technologyare likely to be limited in number and not easyto replace.

Supply Adjustments

Outside of the two sectors described above,there are usually adjustments available to narrowpotential gaps between supply and demand. Themost obvioui response to a "shortage," at leastin the medium term, is for an increased numberof college students to receive training in that field.

Human capital theory predicts that studentswill, on the average, choose the profession thatoffers them the highest return on their investmentin education." Assuming that educational costsat any college do not vary by major, the choiceof major should depend on expected wages andthe ease of advancement.

It is not possible to measure ease of advance-ment, but one can measure wages. Over the past10 years, the starting wage premium that a newlygraduated engineer commands over the averageltewly graduated professional has fluctuated froma low of 7.5 percent in 1975, to a high of 21.1 per-cent in 1981." During this period the number ofengineering baccalaureates increased from 38,000to 63,000. Since a large proportion of employers(35 percent in Massachusetts) adjust their wages

"Gary Becker, Human Capital (New York: National Bureau ofEconomic Research, 196E.

"Douglas Braddock, 'The job Market for Engineers," Occupa-tional Outlook Quarterly, summer 1983.

upward in response to shortages," the recent be-havior of engineering undergraduates can be L.terpreted as a response to a shortage.

Freeman and Hansen" have constructed regres-sion equations, based on data from 1949 to 1981,linking enrollments in undergraduate engineeringprograms to salaries in engineering jobs and inpossible alternative occupations. They find thatengineering enrollments are very responsive to sal-ary changes: A 1-percent increase in real engineer-ing salaries will, by their model, lead to a 2 to4-percent increase in engineering enrollments,

Despite the responsiveness of college studentsto market Opals, the supply of entry-level engi-neers can fail to adjust optimally to shortage prob-lems for a number of reasons. First, universitiesand colleges may not have sufficientresources tomeet student demands for a particular curriculum.This could happen because budget constraints orinstitutional rigidities prevent the university frompaying high enough salaries to attract engineer-ing faculty, or from investing in state-of-the-artequipment. Additionally, the university may havedifficulty evaluating where a new technology isgoing and how long the needs for a particular typeof training will last. Thus, the university may notwant to make an investment in expensive equip-ment or new personnel in the face of uncertaintyover future demand.

Students face much the same problem of hav-ing to judge the long-run value of choosing a par-ticular subfield. Even if a new specialty is neededtoday, students may steer away until it has be-come better established. Or, conversely, studentsmay react too strongly to current reports of short-ages, as that is the only information available.Past student responses to announcements of"shortages" have resulted in surpluses a fc v yearslater.

In all of these cases, both educational institu-tions and students may be acting rationally, giventhe available information. However, the overallresult may not be socially optimal.

"Harrington, et al., op. cit."Richard Freeman and John Hansen, "Forecasting the Changing

Market for College-Trained Workers,"Responsiveness of TrainingInstitutions to Changing Labor Market Demands, Robert E. Tay-lor, et. al. (eds.) (Columbus. OH: National Research inVocational Education, Ohio State University Press, 1983).

t , ('2,

101

Occupational Mobility

An increase in the number of engineering B.S.sin response to a wage increase is not a short-termremedy to a supply-demand gap. It takes 4 ormore years to train a new engineer in college. Ina fast-moving market the needs of employers willhave undoubtedly cha. y the time enteringstudents graduate. TN.. .erg effect is especiallypowerful for occupations subject to sudden un-anticipated surges in demand. The semiconduc-tor boom of the past 1.0 years, for example, was1:irgely unanticipated. How are jobs it such ra.ndgrowth areas filled?

One possibility is that people can move overto these "shortage" jobs from related fields. Alarge amount of independent evidence exists forsuch occupational mobility. An NSF study reportsthat 8 percent of those employed as mathemati-cians in 1972 had become engineers by 1978.20Over the same period, about 5 percent of physi-cal scientists switched into engineering. Overall,approximately one-quarter of all scientists andengineers changed occupations between 1972 and1978.

Attempts have been made to create occupa-tional mobility models on the level of detail nec-essary to project supply and demand. Shaw, etal., created a model in which occupational mo-bility was responsive to wage differences betweenoccupations, as it should be if occupational mo-bility was to eliminate shortages." Moreover, thrmagnitude of effects they found were significantrelative to the size of current shortages.

The bottom line is that occupational mobilityexists, it is responsive to relative demand, and itis important in reducing shortages. It is sugges-tive that when respondents to the 1983 AEA sur-vey were asked how they adjusted to a shortage,almost half reported they would substitute fromother specialties."

"National Science Foundation, "Occupation Mobility of Scien-tists of Engineers," Special Report 80-317, 1980.

"Kathryn 0. Shaw, et al., Interoccupational Mobility of Experi-enced Scientists and Engineyrs, CPA 82-15 (Cambridge, MA: Mas-sachusetts Institute of Technology, Center fo- Policy Alternatives,1982).

"Technical Employment Projects, 1983-87, op. cit.

The major concern about occupational mobil-ity, from the employer's point of view, is that itmay lower the quality of engineering work. Thiswas a major theme of the 1984 NRC symposium.We do not have the direct data to evaluate thisproblem, but we can draw some inferences fromthe patterns of shortages.

First, as noced above, there have not been manyreports of shortages at the entry level in recentyears. Thus, any occupational mobility occurringat the entry levele.g., college graduates whomajored in physics but are taking engineeringpositionshas apparently no, been causing seri-ous problems. This can be explained by notingthat entry-level personnelengineers or other-wisealways have to be trained or supervised.

On the other hand, quality does seem to be aproblem in hiring experienced engineers. This isreflected both in the complaints of shortages ofexperienced engineers, and also in the responsesto direct surveys. Harrington (1983) asked sur-vey respondents what difficulties they encoun-tered in filling job openings for experienced EEengineers. About one-third cited wage competi-tion as the major problem, but the remaining two-thirds of respondents pointed to lack of occupa-tion-specific work experience, lack of industry-specific work experience, and inadequate level oftraining by educational institutions" as contrib-uting to the experienced EE shortage."

Why should occupational mobility lead to morequality problems on the experienced level than onthe entry level? The critical characteristic of anexperienced engineer is that he or she be able towork with relatively little supervision on portionsof a project. Someone switching occupations, bycontrast, will inevitably need additional trainingand experience. It is impossible to turn a physi-cist into an engineer with 5 rs' experience over-night, although experienct: 'Ir. resolving problemsand handling responsibilities in one field can beuseful in anoiher.

The same is true when occupations are definedmore narrowly. A major complaint by firms isthat there is a shortage of experienced electricalengineers with expertise in particular subfields.

"Han ir.gton, et al., op. cit.

..; I 1 Zal

AMIMPN"L

102

The problem is that experienced electrical engi-neers specializing in other subfields have not beenworking for 5 years on the particular technologyrequired by their new employer. Firms willing tohire experienced EEs without the desired skills,and then retrain them, are probably experienc-ing some short-term loss in productivity.

Demand Adjustments

. Employers of engineers have several availableresponses to a potential gap between supply anddemand. The most straightforward response is tocut back or slow down on research and designwork to fit the available labor pool. The short-age disappears because demand for engineers iscut back.

It is hard to know how important this type ofadjustment is, but cutbacks are probably the re-sponse of last resort. The AEA survey reportedthat only 4 percent of respondents would consider"cutting back busine s or reducing projects" in the" . it shortages.

In a competitive economy, a firm forced to cutback on projects because of a shortage of engi-neers could be said to be facing the fact that thereare more profitable uses for engineers in otherfirms. To be more precise, the present value ofthe expected stream of future profits from hiringthis engineer is al parently higher elsewhere.

However, the U 3. economy has large noncom-petitive sectors. In particular, both the defense in-dustry and the educational sector do not base theiroutput decisions strictly on profit motives. More-over, not everyone in the economy shares thesame expectations about the long-term usefulnessof developing a new technology or product.Therefore, it is difficult to judge whether outputforegone by one firm is compensated for by out-put increased by another firm. If shortages are se-vere enough, all firms may have to cut back theiroutput.

Increased Search and Recruitment

A more common response to potential short-ages is to increase search and recruitment efforts.Search and recruitment options open to employersinclude increased advertisements, more direct con-

tact with colleges, use of independent recruiters,and attempts to hire experienced engineers awayfrom other firms. The latter response may involveeither offering higher wages or benefits.

According to one report, the cost of recruitinga high-tech professional (e.g., an engineer) can beas high as his or her annual salary." Thus, theinvestment of firms in search and recruitment isnot insignificant.

Internal Training and Retrainingof Engineers

A common response to the tight engineeringmarket is for firms to take on more of the sor plyburden themselves. In Massachusetts, almost allEE employers offer tuition reimbursement plans,and half of those surveyed had formal in-housetraining programs for engineers.25 According tothe AEA survey, approximately 70 percent of re-spondents, when faced with a shortage, would tryto "retrain or upgrade current employees."

Over the last few years several of the matur-ing computer companies, such as Data GeneralCorp. and Digital Equipment Corp., have adoptedmore systematic training approaches. Faced witha continuing series of shortages, they have shiftedfrom their traditional "buy" strategy to a "make"strategy for obtaining skilled personnel." This hastwo major consequences. First, it helps lower theturnover rate, which, in 1979, was 35 percent forthe electronics industry compared tc, 10 percentin basic manufacturing. Equally important, it in-creases the effective supply of experienced engi-neers. Since experienced engineers are in shortersupply than entry-level engineers, this may be cru-cial in alleviating shortages.

Rearrangement of Jobs

The third employer response to shortages is torearrange jobs and tasks to better utilize the avail-

"Tom Barocci and Paul Cournoyer, "Make or Buy: ComputerProfessional in a Demand-Driven Environment," Massachusetts In-stitute of Technology, Sloan School, October 1982.

"Harrington, et al., op. cit."Paul Cournoyer, "Mobility of Information Systems Personnel,"

Massachusetts Institute of Technology, Ph.D. thesis, 1983; see alsoMichael Mandel, "Labor Shortages in Rapidly-Growing Industries:The Case of Electrical Engineers," Harvard University, January 1984.

1

able skills. It is not clear how prevalent this re-sponse is. About 50 percent of AEA respondentswould react to a shortage by "increasing produc-tivity of currently employed engineers."

On the other hand, Massachusetts companiesdid not see this as a ptissible response to short-ages. Ha, rington reports that "respondents did notsee utilization of existing engineering staff as acontributing cause of shortages in the field . . .

most firms are unwilling to alter their organiza-tional structures to mitigate shortage problems."27These firms may feel that they have already in-creased productivity as much as they can so thatany additional changes would be counterpro-ductive.

"Harrington, et al., op. cit.

103

Firms sometimes appear to willfully avoid or-ganizational change which would increase prof-its in the face of shortages. It is important tounderstand, however, that some institutional ri-gidities which .::ipear to block the best utilizationof engineering talent may in fact be profit-maxi-mizing in the long run. For example, the possi-bility of promotion into prestigious managementjobs may serve as an incentive to get maximumeffect out of young engineers. To keep older engi-neers in engineering jobs could conceivably in-crease the number of engineers in the short run,but it might have the effect of lowering the totalamount of employee effort. Consequently, the re-fusal of many firms to rearrange their job struc-tures in response to shortages is not surprising.

THE CONSEQUENCES OF ADJUSTMENT

To see how well these adjustments to supply-demand gaps are working in the business world,OTA commissioned an in- depth interview studyof the market for EEs in the Boston area. Electz i-cal and electronic engineers were chosen becausethere have been repeated reports of shortages inthose fields, because EEs serve both the defenseand the civilian sectors, and because employersof these engineers are often producers of new high-technology growth products such as computersand electronic equipment. Interviews were con-ducted in the Boston area by OTA contractor Dr.W. Curtiss Priest of Massachusetts Institute ofTechnology (MIT).

The interviews covered 5 categories: large firms(5), small firms (4), engineering school deans andchairmen (4), university placement and employ-ment agencies (4), the Massachusetts High Tech-nology Council Director, and "headhunters" (2).The interviews focused on the availability of theelectrical engineer in the Boston-Route 128 region.

In selecting firms, a variety of firms were cho-sen, reflecting the type of market (defense and ci-vilian) and products. Also, various firms havedifferent reputations .=,-- hiring requirements andcorporate climates T :ypical contact at eachfirm was the vice president of engineering. The

firms contacted were: Raytheon Corp., AVCOCorp., Wang Laboratories, General Electric (GE),Polaroid Corp., Pacer Systems, Inc., Baird Corp.,Technical Alternatives, Inc., and Apollo Com-puter Corp.

Engineering school deans and chairmen werecontacted at MIT, Northeastern University, TuftsUniversity, and the Franklin Institute. Universityplacement contacts were at MIT and Northeast-ern. Employment agencies contacted includedComputer Placement Unlimited, Inc., High Tech-nology Placement of Winter, Wyman & Co., For-tune Personnel, and McKiernan Associates.

Nearly all of the interviews were made in per-son and lasted from 1 to 2 hours in length. Anopen-ended discussion was conducted with eachcontact using seven categories of questions. Thesequestions investigated how shortages of engineersvary over the short and long term, the availabil-ity of particularly needed skills, and the abilityof engineers to grow into job requirements. Ad-ditionally, the discussion covered the impact ofmilitary procurements, the benefits and environ-ments required to attract engineers, and the sen-sitivity of universities to market demands. Finally,interviewees were asked to compose a "wish list"in order to identify major concerns about the

1 - *"

.t .1.-L

104

availability of engineers. To explore the hiring re-quirements in detail, the respondent was askedto describe a job description and how closely ahire would have to match the description.

Variability in Shortages of Engit.eers

Each firm had a uniquely different set of ex-periences regarding shortages of engine sr.. Theseexperiences depended on the character of the firm,its location, and its reputation. In general, indus-try responded that there were certainly shortagesof some types of engineers at particular times butsome firms reported no shortages because of heirparticular position in the market.

Shortages of engineers vary by geographic loca-tion. While the focus of the study was on theBoston-128 region, a number of firms had multiple locations across the country. Raytheon Corp.has experienced some shortages in the Boston re-gion but extreme shortages in the Santa Barbaraarea. Likewise, Pacer Systems has experiencedsome shortages in the Boston area but more short-ages at their Pennsylvania location where thereis a higher demand for engineers with subs:an-tial military service experience. A number ofplacement respondents commented on the exis-tence of shortages in the Silicon Valley area be-cause of the very high costs of living.

Shortages vary by the firm's reputation and cul-ture. Apollo Computer and Wang Laboratoriesstated that they do no: suffer from any shortagesof engineers. Both firms are successful computerapplications firms with distinctive corporate cli-mates and reputations. Apollo Computer has aunique reputation for fast growth and the "Apolloculture." This places the firm in an "unreal,dreamworld existence" in which they can be ex-tremely selective about applicants. In contrasi, acompany like Baird, an instrumentation firm,often finds it difficult to hire engineers within sal-ary ranges :hey can afford. They prefer to hirefresh graduates to help keep down salary costs.

Shortages vary across general areas related toelectrical engineering. Electrical power generationand distribution is an area that depends on theconstruction market and the demand for electri-city. The demand for electrical engineers in this

area is fairly steady and the supply appears ade-quate. The electronics industry, in contrast hashad :,ubstantial periods of growth, creating a highdemand for the electronics-oriented electrical engi-neer. There have often been periods of shortagesfor particular skill requirements in the electronicsfield. Computer programmers in electrical engi-neering applications areas have also been in heavydemand. However, there is some indication that2-year educational programs are successful in meet-ing the demand for many of these requirements.

Shortages also vary considerably by special skillneeds. There are Lertain positions that requireboth excellent electronics hardware skills andstrong computer programming skills, such as inthe design of the next generation of computers.Qualified people are very scarce. Another currentarea of shortage is in the use of autor..atic testequipment (ATE). Knowledge of ATE has becomequite specialized to certain machines and softwarepackages. It is difficult to locate people with therequired background. Another scarce skill areainvolves engineers with both analog and digitalhardware experience. Many older engineers skilledin analog hardware find it difficult to make thetransition to digital hardware. Younger engineers,more skilled in digital work, find analog hardwaredifficult and foreign.

Changes in markets and opportunities are cons-tantly producing specific skill shortages. The re-cent emphasis on "automatics" such as roboticsand other industrial control machinery has cre-ated a heavy demand for industrial electrical engi-neers. Until a few years ago, industrial engineerswere in low status positions and many universi-ties, such as MIT, did not train industrial engi-neers. Another recent growth area, "ve:y large-scale integrated circuits," has created shortagesof engineers both in industry and universities.

Shortages vary over time as economic condi-tions and growth opportunities change. Respond-ents with 20 to 30 years of experience recall theebb and flow of hiring over a number of cyclesof high and low demand periods. Memories wererecalled of how in the late 1960s, half of the em-ployment agencies went out of business, and sto-ries were circulating of engineers driving taxis.More recently, shortages in 1979 and 1980 were

105

remembered in terms of prolonged searches andless than optimal hires. In prior decades, militaryemployment was viewed as unstable and compa-nies such as Polaroid were considered stable andsecure. In the current decade, military employ-ment is seen as a refuge and the frequent cyclesof civilian layoffs are viewed with greater anxi-ety. The current softening of the market is viewedby many as a temporary adjustment period.

Shortages are not perceived to vary much bythe level of military procurements. Many respond-ents emphasized that the two climates of militaryand civilian work are so different that there is lowmobility between the two "sectors." The militaryengineer is viewed as more risk averse, less crea-tive, and less likely to be interested in advance-ment. In contrast, the civilian engineer is viewedas more people oriented, more talented, betterable to bring out products, and more selective.It is more likely that engineers will move into themilitary s. " or when civilian hiring is down thanwill military engineers move Lie) the civilian sec-tor when hiring in the military sector is down.Thus, when military procurements are rapidly in-creasing, those hiring in the civilian sector senseless competition in hiring the same people thanthey might if the mobility were higher.

Raytheon, a company with about 75 percentof its engineers conducting military wot k, indi-cated that they keep salaries at about the samelevel for both civilian and military positions re-gardless of changes in demand for m;','Itary work.Pacer Systems indicated that tight government au-diting practices kept them from bidding up sala-ries for military engineers. While some of thesestatements appear to defy the economic laws ofsupply and demand, they were prevalent enoughto warrant further investigation.

Impact of Shortages

In general, shortages tend to lengthen searchtimes and may relax, somewhat, the requirementsrequired for a position. Many respondents statedthat, for the most part, the supply of engineershad improved over the last 5 to 8 years. Even dur-ing high demand periods, the impact of shortageshas not been dramatic.

Changes in the level of shortages of electricalengineers substantially -affects the search time fora new hire. Responses about variability in searchtime were fairly consistent. During periods of rea-sonable availability the search period takes 1 to11/2 months. During periods of greater shortage,this time can easily double and a search periodof 4 to 5 months would not be unusual. In theseperiods, lesser known firms must be prepared tosettle for less desired choices in candidates. Spe-cial needs require more search time; the processof hiring the right chief engineer when the require-ments are fairly unique can take more than a year.

Shortages have an impact on the quality ofengineers and productivity. For example, duringthe recent 1979-80 shortage, a number of respond-ents said that significant compromises were madein hiring. The level of experience criterion was metless often. Also, firms used to hiring from "firstrate" schools might find it necessary to relax theirrequirem its to include other schools during highshortagr tionetheless, there was a sense that nolarge , ures from expectations were evermade. Tiie compromises were significant but nothighly burdensome.

Shortages also have an impact on deferred op-portunities. For example, new hires in specificareas may add to the future research capabilitiesof a firm, but a shortage in that area presents an"opportunity cost" to the company's R&D effort.Polaroid encountered shortages of Ph.D. opticalengineers and AVCO described shortages of "hardscience" based engineers that impeded expansionof their research capabilities.

A major impact of shortages may be job-switching and high expectations of engineers. A"star" engineer (heavily sought after for his or herhigh abilities and relevant skill area) will oftenswitch jobs every year or two. The change is typi-cally made to receive more responsibilities, workon more challenging problems, and for the cor-responding :.glary increase.

The electrical engineer has fairly high expecta-tions about the job in terms of challenge, goodrelationships with the supervisor, promise of R&Dwork, freedom and independence, cutting edgeprojects, being entrepreneurial in design, and high

106

salary. With a very high degree of consistency,respondents described these expectations and theimportance of meeting them to hire and retain theengineer. It was also emphasized that a good sal-ary went along with the other expectations andwas not typically sought after as the primary goal.

Mobility and Flexibility of Engineersand Related Profess;onals

In general, it is difficult for an engineer to takea job requiring work much different than his orher previot s experience. There are certain "feeder"field; like math and physics but it is nearly im-post .ble for, say, a chemical engineer to take aposition as an electrical engineer, or even for anelectr;cal engineer experienced in radio-frequencydesign to take a job in computer design. Profes-sionals in math and physics are often hired for"system.;" work because their general training isuseful in general problem-solving and design.They are more likely to be hired by firms engagedin basic research, such as Polaroid and AVCO.They are less likely to be hired by a firm likeApollo Computer. Thus, when the demand. ishigher for engineers than scientists, as has beenthe case over the last 10 years, mathematiciansand physiists can shift to fill some of the demand.To illustrate the magnitude of this shift, from 1973to 1983 the number of undergraduates in engineer-ing at MIT (primarily electrical engineers) Ire,gone from 1,257 to 2,405. Over the same timeperiod, the number of undergraduates in theschool of science has declined from 1,162 to 725.

Not all engineers are qually "flexible" or "mo-bile." The engineer in the top quartile of hisprofession is considered much more mobile andflexible than the other 75 percent in the profes-sion. Engineers that have reached positions ofmanagement or high rank typically havc greaterflexibility and mobility than -new.: junior engi-neers. Three factors seem to play a part in restrict-ing the flexibility and mobility of engineersmarket limitations, management restrictions, andknow-how limitations. These factors are not in-dependent, but work together to reduce mobil-ity and, at the extreme, to render obsolescentmany tagineers.

In terms of market limitations, firms oftenchange while in a period of crisis. They requireparticular engineering skills immediately and ob-tain these skills from the outside market becauseit not pr'ctical to retrain in-house engineers inthe time aval5.ble. In terms of management re-strictions, engineers become identified by man-agement as having certain abilities and know-how; it is difficult for management to take therisks associated with allowing an engineer to makea transition to a new skill area.

In terms of know-how, engineering is a highly"know-how" intensive profession. It takes con-siderable training and practice to be proficient ina particular area of electrical engineering andwhile some of these skills are transferable, manyare not. For example, analog circuitry know-howis not highly transferable to digital circuitry. Thus,over the last 20 years when the field was movingfrom analog to digital, many engineers becameincreasingly outmoded.

The interviews constantly reinforced this dy-namic picture of the way many c.gineers becomeunable to face periods of transition. Many re-spondents spoke of the need for more engineersto "keep up." There were stories about the keyengineer who did keep up and how he or sheplayed a signifiant role in the newer areas, in con-trast to those who did not. The placement offi-cer at Northeastern referred to those engineerswho worked in the same area for 10 to 15 yearsand then had to return to drafting or some otheroccupation requiring less skills when the areadried up. While many professional fields are facedwith technological change, the engineering profes-sion is probably more vulnerable than most.

For the more mobile and flexible engineers, theheadhunter plays a critical and dynamic role. Theemployment headhunter is a placement personwho actively seeks out individuals to meet aclient's employment needs. Many employmenta3encies receive applications from engineers andmatch these with employment opportunities theyhave identified in industry. The headhunter goesone stet- fTrther and attempts to lure a "star" engi-neer out of one company to another. As disrup-tive as this first appears, the headhunter plays an

1 4..1. II LI

107

important role from the standpoint of economicefficiency. He or she is helping to allocate a scarceresource to a more optimal allocation. The head-hunter establishes a network of contacts bothwithin organizations looking for talent and withinorganizations where the talent is "misapplied" or"underutilized."

In matching individuals, the headhunter is sen-sitive to the flexibility of the employer in termsof the technical capabilities of the individual andthe personal match. Often it was said that em-ployers hire "people they like." Thus engineersfrom certain styles and cultures of firms are muchmore likely to fit a particular position than others.But other placement professionals warned againstoveremphasizing the cultural variable. The bot-tom line is how well the person can do the joband as monolithic as many firms appear, thereare many subcultures within any given orga-nization.

In summary, many factors play a role in re-stricting mobility and flexibility. These relate tobc th technical variables and cultural variables.For small incremental shifts in the work force,these factors will not be highly significant. For ma-jor occurrences of technological change and largechanges in the economy, however, these factorswill be quite important.

Desirable Changes

In concluding each interview, participants wereasked to identify the most desirable changes re-lating to the availability of engineers. The re-sponses addressed various availability, quality,and professional issues.

In terms of shortages, people in employmentagencies all wished they could have a larger. morecapable pool of engineers to draw from. Sincethey are personally responsible for how well theymeet their clients' expectations, this is a signifi-cant statement about the shortage of engineers.Within industry, if a firm required a particularspecialty which was in short supply, the responsewas to have more engineers trained in that spe-cialty. For example, the GE generators sectionwished more schools provided "power systems

engineers." Raytheon, involved in more radiowork, wished schools trained more radio-fre-quency engineers. In many cases, the undersupply was in a low glamour area. Power systemsengineers, industrial engineers, etc., are lower sta-tus areas and many students prefer the high tech,high glamour areas. Whccher "glamour" is a "cor-rect" economic market signal for whether studentsshould pursue a particular area of electrical engi-neering is an interesting question.

This leads to a very serious issue of the statusof engineers in general. Respondents from univer-sities unanimously indicated that they were up-set about the "second class" treatment of engi-neers. One respondent described industry astreating engineers as a commodity while anotherindicated that engineers are often used for "sub-professional tasks." The dean of one engineeringschool suggested that the supply of engineersshould be limited, as the medical profession hasdone with doctors, so that the standards of profes-sional excellence and stature could be raised. TheGerman model was considered a success in main-taining engineering professional standards. Theremay be a linkage between the status issue and thefactors that lead to immobility and inflexibilitydiscussed above. How can an engineer be a profes-sional if so much of his or her know-how can bebe made obsolete by the advancement of technol-ogy? One response to this dilemma has been toinstitute lifelong learning."

Lifelong learning was emphasized by both re-spondents in industry and in the universities. Arecent conference on "Keeping Pace With Change:The Challenge for Engineers" was sponsored bythe Massachusetts High Technology Council andheld at Northeastern University. In a study spon-sored by the High Tech Council, the "half-life"of the engineering degree (length of time half ofthe knowledge acquired by graduation remainsrelevant to work tasks) was estimated to be be-tween 3 and 5 years for the Bachelor of Sciencein Electrical Engineering. The peak age for the per-formance of engineers, in terms of age, was foundto be the late twenties and early thirties. Accord-ing to this study, productivity typically declinesfrom age 30 until retirement.

108

To counteract this process, lifelong learningprograms have been established within industriesand through universities. It was noted that theengineer over age 35 cannot be expected to returnto the classroom but nee& different educationalresources to keep up. The older engineer needssupport groups in the company of his peers. Oneprogram is provided by Northeastern Universitywhere live graduate courses are provided usingvideo to locations along Route 128 in Boston. Thesystem has two-way voice for communication andcourier service for homework.

Perhaps, too, with newer information technol-ogies, the engineer can keep pace by greater useof information databases where the knowledge iskeyed by function and application. While generalcitation databases exist such as COMPENDEX onDialog Information Services, these only providebibliographic citations to the engineering litera-turea literature which is itself surprisingly /owon useful material. If the private market cannotrespond in providing these information services,it may be a role of government to help meet thisneed.

Respondents also expressed the wish for engi-neers to have additional skills beyond basic engi-neering capabilities. A few respondents empha-

sized the need for engineers to be better able t.)communicate and write. A few indicated the needfor engineers to be better able to use computersimulation to test a prototype design in place ofbuilding the prototype because of the high costsinvolved in making custom integrated circuits.Two respondents emphasized the need for engi-neers to have stronger skills in scheduling andplanning, and a better sense of management's ob-jectives.

Finally, Polaroid, AVCO, and Raytheon indi-cated the need for government and industry toprovide greater support of hard science and ap-plied research. A vice president at AVCO washighly concerned about the loss of hard sciencePh.D.s since the Vietnam W r who could not onlycontribute to military strength but to basic re-search vital to civilian industrial strength. Fromthe university side, Tufts noted that the currentNSF emphasis on large research programs hasgreatly reduced opportunities to support researchat Tufts. The Commonwealth of Massachusetts'program to provide research facilities in micro-electronics available to universities and industrywas lauded as a way of providing facilities thatTufts could use to educate their engineeringstudents.

CONCLUSION

Shortages exist for engineers, especially forthose in a few high demand fields. In the last 5to 8 years, however, shortages have been trouble-some but not a large burden. Some firms havenever perceived shortages while others find thesupply of engineers much too small. The flexibil-ity and mobility of some engineers is high but formost engineers it is low. The engineering marketperforms reasonably well for gradual transitionsbut does not perform well during periods of ma-jor technological change or dramatic changes inmarket and/or economic conditions.

In general. there does not appear to be a needfor a direct government role in alleviating poten-tial shortages of engineers. The market for entry-level engineers seems to be functioning well enoughnow to eliminate shortages. At the bachelor's

level, some educational institutions report thatthey do not have enough resources to train all theengineers that are interested, due to shortages offaculty and equipment.

Although there have been, and will continueto be. spot shortages of experienced engineers,they are not of a type amenable to direct govern-ment intervention. When a new technology is de-veloped, it simply takes time before there is a poolof engineers who are experienced with the newtechniques. There may be a role for the govern-ment in helping to promote the retraining of ex-perienced engineers into startup fields. Mission-agency sponsored R&D programs could conceiv-ably serve that retraining function. The Federalenergy and environment programs of the 1970s,for example, helped retrain signifscart numbers

109

of scientists and engineers to become specialistsin those two fields. Many of these people are cur-rently employed in their new capacity by in-dustry.

A fundamental problem is the obsolescence ofknow-how and the corresponding obsolescenceof engineers. In response, many institutions andfirms are encouraging lifelong learning. Since gov-ernment has traditionally played an importantrole in education, there are some policy optionsw!..!c..% may help. One important role is to pro-vide basic facilities for education such as themicroelectronic center that Tufts and other univer-sities find important in helping train engineers.Another role is to identify economic disincentivesto lifelong learning and provide either educationalsupport or tax relief to enable more engineers tokeep up their know-how.

In the area of information policy, there is thequestion of how new information technologycan

1

*

assist in providing ongoing know-how supportand lifelong learning to engineers. Does the struc-ture of the market of engineers and their em-ployers encourage the private sector developmentof such information support systems, or are therepublic sector reasons for government support? Anumber of market failure reasons can exist for thelack of the development of such support systems.To the extent that firms do not fully bear the costsof the obsolete engineer, firms will underinvestin further training, etc. Also, information hastransferability problemsit is often difficult forthe supplier to exact the full value of the infor-mation because it is difficult to keep the informa-tion from being passed along (with no further rev-enue to its producer).

"For a comprehensive discussion of these and other reacons seeW. Curtiss Priest, "The Character of Information," contractor pa-per for U.S. C digress, Office of Technology Assessment, Project -on Intellectual property, Feb. 10, 1965.

Chapter 5

Demographics andEquality of Opportunity

c.i

Contents

PageWomen . . . . 114

The Science and Engineering "Pipeline" . 114Differential Treatment of Women in the Science and Engineering Work Force 118Implications for Women's Participation in Science and Engineering 121

Minorities . . 122Differential Treatment in the Work Force . . 126

Effectiveness .I Programs to Promote Participation 127Policy Implications . . ..... . 128

List of TablesTable No Page5-1. Average Annual Salaries of Scientists and Engineers by Field and

Sex/Race/Ethnic Group, 1982.... ..... . . . 1195-2. 1978/79 Representation in Quantitatively Based Fields Relative to Representation

in Age-Relevant Population by Degree Level and Racial and Ethnic Group . 1225-3. 1978/79 Representation Relative to Representation in the

Age-Relevant Population by Degree Level and Racial and Ethnic Group . 1225-4. 1978/79 Representation in Quantitatively Based Fields Relative o Representation

in Total Deg-ees by Degree Level and Racial and Ethnic Group 122

FigureFigure No Page5 1. Percent of Degrees Granted to Women 1950-82 . 115

Chapter 5

Demographics andEquality of Opportunity

The principle of equality of opportunity inscientific and engineering careers is embodied inthe Equality of Opportunities in Science and Tech-nology Act of 1981:'

The Congress declares it is the policy of theUnited States to encourage men and women,equally, of all ethnic, racial and economic back-grounds to acquire skills in science and mathe-matics, to have equal opportunity in education,training, and employment in scientific and tech-nical fields, and thereby to promote scientific liter-acy and the full use of the human resources of theNation in science and technology. To this end,the Congress declares that the highest qualityscience over the long-term requires substantialsupport, from currently available research andeducational funds, for increased participation inscience and technology by women and minorities.

The Act sets out three reasons for promotingequality of opportunity: equity, the need for gen-eral scientific literacy, and the desire to fully uti-lize all human resources that could be applied toscience and technology. It also explicitly calls forthe commitment of Federal resources, within theoverall scientific research and training budget, toprograms which promote increased participationby minorities and women in science and technol-ogy. This commitment has not been completelycarried out.

The changing college student demographics ofthe coming decade have several important impli-cations for national policy regarding equality ofopportunity. If the number of college graduatesdeclines substantially, as predicted, the principalof utilization of all potential human resources tothe fullest extent possible will become especiallyimportant. Although, as argued in chapter 1, itcannot be proven conclusively that the Nation re-quires the current level of 300,000 new science andengineering bachelor's degree recipients per year,it would not appear prudent, in an .-ra when

'National Science Foundation AuthJrization Act of 1981. PublicLaw 96-516, Sectiol 32(b).

science and technology are expected by many toplay an increasing role in improving our economiccompetitiveness and national security, to allowthe number to decline appreciably. Therefore,from a "utilization of resources" point of view,programs to promote participation of historicallydisadvantaged groups take on greater significance.

The increasing fraction of college students whoare likely to be drawn from the black and Hispani.:populations, that have historically participated inscience and engineering education and employ-ment at far lower rates than the white population,imply that programs clirect^d at these two minor-ity groups may be needed to keep the supply ofscientists and engineers from declining.

In order to understand what will be required,it is important first to have a clear picture of thefactors leading to reduced participation by womenand certain minority groups in science and engi-neering. These factors are complex, interrelated,and span the full developmental cycle from earlychildhood socialization to experiences in the workforce. They include:

the continuing legacy of decades of discrimi-nation and discouragement from scientificand engineering careers,differential treatment of women and minor-ities in the science and engineering workforce,lack of early educational opportunities dueto social class and cultural factors among mi-norities,female socialization patterns that discourageyoung women from perceived "masculine"careers,expectations that women will continue to as-sume the major role in housekeeping andchildrearing and sacrifice their professionalinterests to those of their husband,lack of financial support and institutionalbiases in the higher education system, andlack of role models and early exposure to

0113

114

science and engineering as worthwhile andaccessible careers.

The sections that follow will review, briefly,the participation rates in science and engineeringeducation and employment for both women andminorities, and discuss in some detail the prin-dpal causes, as currently understood, of those lowparticipation rates. The limited evidence on theeffectiveness of intervention rrograms that have

been initiated to increase participation in scienceand engineering among these groups will be pre-sented. Finally, a set of policy and reseamn issuesrelated to increasing the participation of womenand minorities in science and engineering will begiven. These issues have been identified by a re-view of the literature, an OTA workshop, anda questionnaire to practitioners in the field.

WOMEN

The increasing participation of women in scien-tific and engineering education and employmentis a well-documented phenomenon. The trend ismost dramatic in higher education, where women'sshare of total science and engineering baccalaure-ates increased from 28 to 36 percent; of master'sdegrees from 18 to 28 percent; and of doctoratesfrom 15.5 to 29.4 percent between 1972 and 1982.Figure 5-12 illustrates that increasing participationby field and degree. In employment the trend isless dramatic but still significant. Women con-stituted 8 percent of the Nation's scientific andengineering work force in 1973; a decade laterthey were 13.1 percent. The most impressive gainswere in computer specialists, engineers, and lifescientists, where the numbers of women increasedby factors of 4, 3, and 2 respectively between 1976and 1983.3

Two issues link women's increasing participa-tion in science and engineering to the demographictrends which are the subject of :his technicalmemorandum:

1. the degree to which the increasing partici-pation of women in science and engineeringis likely to continue over the next two dec-ades, and

2. the implications of overall demographictrends for policies and programs to promoteequality of opportunity for women in scienceand engineering.

'Betty M. Vetter and Eleanor L. Babco. Professional Women andMinorities, 5th ed. (Washington, DC: Scientific Manpower Com-mission, August 1984), p. 37.

'Science and Engineering Personnel: A National Overview, NSF85-302 (Washington, DC: National Science Foundation, 1985), pp.53-54.

To understand these issues better, it is neces-sary to examine the factors that influence women'sdecisions to enter and remain in science and engi-neering careers.

The Science and Engineering"Pipeline"

According to Sue Berryman, the pool of talentfrom which the Nation's scientists and engineersis drawn is largely formed in high school. Inter-est in science and mathematics first appears inelementary school, develops most intensely priorto 9th grade, and basically is completed by the12th grade. After high school, migration is almostentirely out of, not into, the pool. As a conse-quence:

. . . those who obtain quantitative doctorates orhave mathematically-oriented careers a decade af-ter high school come overwhelmingly fromgroup in grade 12 who had scientific and mathe-matical career interests and high mathematicalachievement scores . . . . By grade 12, theseachievement scores clearly differentiate those whoplan to attend] college from those who do notand those who plan quantitative college majorsfrom those who plan non-quantitative ones.'

II we follow a group of 4,000 seventh gradershalf boys and half girlsthrough the sequenceof steps from age 12 through 32 that ultimatelyselect those who will become scientists and engi-neers in quantitative fields (the physical sciences,mathematical and computer sciences, biologicalsciences, economics and engineering), we find the

'Sue E. Berryman Whn Will Do Science (New York: The Rocke-feller Foundation, November 1983), pp. 66-77.

115

I

Figure 54.Percent of Degrees Granted to Women 1950-82

1960 1980 1970

Year

1980 1950 1900 19'0

Year

1980.

SOURCE: Scientific Manpower Commission, Profresionel Won,en and Minorities (Washington, DC 1984), p 38

following interesting patterns.' About half eachof the boys and girls will have taken and under-stood enough mathematics by age 12 to be capa-ble of pursuing sufficient advanced work in highschool to prepare for a quantitative n.ajor in col-legt. (At age 12 the mathematical abilities of boysand girls are almost identical.) Of this group of1,000 boys and an equal number of girls, only 280boys and 220 girls will actually take enough highschool mathematics to major in a quantitativefield in college. This difference in numbers of stu-dents taking advanced mathematics courses is oneof the factors leading to the observed differences

'Betty M. Vetter, 'The Science and Engineering Talent Pool" inScientific Manpower Commission, Proceedings of the 1984 JointMeeting of the Scientific Manpower Commission and the EngineeringManpower Commission (Washington, DC: National Academy ofSciences, May 1984), pp. 3 and 5.

1950 1980 1970

Year

1980

between young men and young women in collegeSAT mathematics scores.

Of the original 2,000 young potential scientistsand engineers, only 140 of the boys and 45 of thegirls will actually enter college with plans to ma-jor in science or engineering. Forty-five of thoseyoung men will emerge from college with a quan-titative baccalaureate degree; 20 of the women willdo likewise. At the Ph.D. level, five men and onewoman, of the original science and engineeringpool of 2,000, will actually receive a doctoratein a quantitative field.

We can see from the above statistics just wherewomen's participation in quantitative fields dropsoff most sharply. Between the age of 12 and theselection of a potential major in college at age 18,young women's persistence in the science and

122

118

engineering "talent pool" falls off three times asrapidly as that of young men. After college, 1 outof every 20 women, as opposed to 1 out of every9 men, who receive quantitative 13.A.s goes onto complete a quantitative Ph.D. Thus, it is at thepre-college and postgraduate levels that we shouldlook for factors that discourage women from pur-suing science and engineering careers.

The factors that lead college freshmen womento pursue science and engineering majors less fr--quently than men are best illustrated by a com-parison of major field preferences among collegefreshmen. The 1984 study of T Y American Fresh-man by the Higher Education Aesearch Institute'shows exactly where the two sexes differ most intheir preferences. Twenty percent of the men, butonly 3 percent of the women, report an intentionto major in engineering. Three and one-third per-cent of the men, but only 2 percent of the womenreport an intention to major in a physical science.By contrast, women show significantly greaterpreference ror the social and the biological sci-ences; 8.4 to 5.1 percent in the social sciences; 4.5to 4.1 percent in the biological sciences.

These differences are often explained as conse-quences of ..oung women's lack of exposure to,or poorer performance in high school mathe-matics. However, The Am rican Freshman sur-vey and the records of earned baccalaureates castconsiderable doubt on this explanation. Accord-ing to the survey, 0.9 percent of the freshmenwomen and 0.8 percent of the freshmen men sur-veyed in 1984 intended to major in mathematics.More than 7 percent of the women and only 5.7percent of the me :ntended to specialize in ac-cot nting. Twenty-five percent of the men and 23pea -..ent of the womim intended to major in a busi-ness field other than "secretarial studies." Accord-ing to the Scientific Manpower Commission,'more than 40 percent of the bachelor's degrees inmathematics have been received by women everyyear since 1974. Close to 40 percent of the busi-ness and management degrees and close to 35percent of the computer science degrees were

Alexander W. Arlin, et al., The American Freshman: NationalNess for Fall 1984 (Los Angeles, CA: The Cooperative Institu-tional Research Institute, University of California at Los A.igeles,December 1984), pp. 16-18, 32-34.

'Vetter and Babco, op. tit., p. 37.

awarded to women in 1982. Thus, it does not ap-pear that women are avoiding quantitative fieldsin general due to perceived inadequacies in theirmathematical capabilities. Rather, it appears thatwomen are avoiding engineering in particulardespite significant increases in their enrollment inthat field in recent yearsand, to a lesser degree,physics.

To understand why women tend to select engi-neering as a college major so much less frequentlythan men do, it is useful to examine those fieldswhich women tend to choose far more often thanmen. The fields with the largest differentials infavor of women in The American Freshman are:education (9.6 percent of the women, 2.8 percentof the men); nursing (7.6 percent of the women,0.3 percent of the men); and "ierapy (3.4 percentof the women, 0.8 percent of the men). Thosethree fields account for 20.6 percent of the womenand 3.9 percent of the men in the survey, a mir-ror image of the ratio in engineering (3.0 percentof the women, 20.1 percent of the men).

Education, nursing, and therapy are all tradi-tional "women's fields," while engineering is tradi-tionally a "man's field." Education, nursing, andtherapy are all considered to be "helping" and"people-oriented" professions, whereas engineer-ing is associated with building things and control-mg the physical universe. Freshmen women's

tendency to select the first three fields over eng'-neering may simply indicate that traditional sex-role stereotypes and career patterns have not yetworked their way out of the system.

This should hardly be surprising from a histori-cal point of view. '''; is less than a decade and ahalf since the women's movement forced open thedoors to the traditionally male professions. Allof the parents and most of the teachers of the cur-rent crop of freshmen were raised in the pre-feminist era. There are significant groups andleaders in the country who are not fully commit-ted to the ideals of equality between the sexes inthe work force. Thus, the influential adults whohelp to shape teenagers' ideas about the world andtheir place in it are not uniformly aligned in fa-vor of efforts to break with traditional sex rolesand occupational choices. The persistence of a siz-able fraction of college freshmen who rem: in true

4- kii'1

-..

to earlier stereotypes of "women's" and "men's"work in these circumstances is hardly surprising.

Recent research by Gail Thomas tends to con-firm the persistence of traditional sex-role stereo-types. Thomas found, from an extended study of"Determinants and Motivations Underlying theCollege Major Choice of Race and Sex Groups,"that the "choice of a college major entails a fairlypredictable process that is largely formalized priorto college" and is "characterized by distinct (andto a great ext,at traditional) male and femaleinterests, values, and aspirations held duringchildhood." These findings, according to Thomas,"imply the importance of traditional sex social-ization. . . ."8

Sue Berryman finds an additional career-relatedfactor helping to explain young women's decisionsto avoid quantitative majors in college. Thegreater a woman's expectation to assume the ma-jor child rearing responsibilities of her children,the less likely she is to choose quantitative occu-pations that require major educational and laborforce commitments. The more she expects con-tinuous labor force participation during adult-hood, the more her occupational goals approxi-mate those of young men. Berryman notes that"studies show that male single parents m '.;e oc-cupational and labor force adaptations to parent-ing that are similar to the plans of young womenwho expect dual family and work responsibili-ties."'

Berryman sees young women's career expecta-tions interacting with their decision to takeadvanced mathematics courses to produce the pat-tern of underrepresentation of women in quan-titative fields. "Gender differences in grade 12mathematics achievement are primarily attributa-ble to differences in boys' and girls' participationin elective mathematics during the 4 years in sen-ior high school," according to Berryman. Priorto grade 9, boys and girls do not differ signifi-cantly in average mathematical achievement. The

'Gail E. Thomas, Determinants and Motivations Underlying theCollege Major Choice of Race and Sex Groups (Baltimore, MD. Cen-ter for Social Organizations of Schools, Johns Hopkins University,March 1983), p. 43.

'Sue E. Berryman, "Minorities and Women in Mathematics andScience: Who Chooses These Fields and Why7" presented at the 1985annual meetings of the American As an for the Advancementof Science in Los Angeles, CA, Ma 15, p. 14.

117

individual's confidence in his or her mathematicsability, and perception of the utility mathematicswill play in achieving educational and careergoals, are factors contributing to the participa-tion in the high school mathematics sequence. Thestronger the two factors are, the greater the likeli-hood of partipation. Since career goals CP'M todetermine educational investments, gende. Differ-ences in occupational expectations become key tounderstanding gender differences in high schoolmathematics participation. Berryman sums up bystating: "the key for women seems to be their ca-reer choices, their investment in the junior an!senior high school mathematics and science se-quences being related to these choices" (empha-sis added).10

Although Berryman sees mathematics abilityand achievement as crucial to success in a quan-titative career ("high mathematical achievementat grade 12 predicts realization of grade 12 quan-titative career plans by age 29"), she sees suchachievement as strongly related to, and influencedby, career choices already being formed in highschool.

Once in college, it appears that young womenhave significantly less trouble than young mencompleting a quantitative baccalaureate. Forty-two percent of the women, as opposed to 30 per-cent of the men, who begin college with a quan-titative major emerge after 4 years with a quan-titative B.A. In graduate school, however, furtherattrition occurs. Women receive less than one-third as may quantitative master's degrees andone-fifth as many Ph.D.s. As a fraction of quan-titative B.A.s women's attairment of quantitativeM.A.s and Ph.D.s is less than half that of men.

The causes of women's attrition from the scienceand engineering "tale nool" in graduate schoolare not well documented. The problem does notappear to be with initial enrollment. The percent-age of science and engineering graduate studentswho are women is almost identical to the percent-age of science and engineering B.A.s awarded towomen. There are some differences from field tofield, with a considerable underrepresentation ofwomen in mathematics and a significant over-representation in life science graduate programs,

' °ibid., pp. 13-14.

1 7: 1. l 4 - -lt

but overall the numbers are quite close, as canbe seen below:"

Reid Percent B.A.,women, 1980-82

Percent grad. enroll.,women, 1982

Total, S/E 36 .6% 35.3%Physical sciences 24 .5 21.0Engineering .11 .0 10.9Mathematic a. sciences... .36 .9 27.1Life sciences 40. 0 52.4Social sciences 51. 8 47.0

The problem appears to be with persistence ingraduate school, with womer's participation de-creasing "at each successive degree level."" ShirleyMalcom cites one possible cause of this declinelack of financial support:"

In the 1981 Summary Report of Doctorate Re-cipients, a discussion of differences in financialsupport of doctoral training is a cause for seriousconcern. Women were more likely to report "self"sources of support. This was the primary sourceof support for 45 percent of the women but only30 percent of the men. On the other hand, re-search assistantships were reported as the primarysource of support for the doctorate by over twiceas many men (22 percent) as women (10 percent).

Since research assistantships facilitate entry intoa research career, by providing access to equip-ment, mentors, conferences, and publications, thedifferential access to research assistant support ap-pears to be an especially important problem.

A second factor, however, is undoubtedlywomen graduate students' increasingly negativeperceptions of the act ial benefits their advanceddegree training will bring them. Once in gradu-ate school, women can often see in the behaviorsof their professors the forms of discriminationthey will face in the workplace, reducing consider-ably the return on their investment in higher edu-cation.

"Vetter and Babco, op. cit., pp. 27-37."Shirley M. Malcom, Women in Science and Engineering An

Overview, prepared for the National Academy of Sciences (Wash-ington, DC: American Association for the Advancement of Science,September 1983), pp. 27 and 37.

p. 8.

Differential Treatment sWomen in the Science andEngineering Work Force

It would be nice to report that once a womanhas gone through the time and trouble to receivetraining in a science or engineering field, especiallyto the Ph.D. level, she is safely ensconced in thetechnical work force with as great a likelihood ofremaining there and receiving the full benefits ofher education as tier male counterpart. Unfortu-nately that is not the case. Sue Berryman findsthat "labor force attrition rates differ far more bygender than by race" with female attrition rates'more than 50 percent higher than those of men."

In 1982, almost 25 percent of the women trainedas scientists and engineers, compared to 16 per-cent of the men, were not using their training inthe scientific and engineering labor force. One-third of the women who were out of the laborforce in 1982 had left for reasons of family respon-sibilities."

Lilli Hornig, in an article entitled "Women inScience and Engineering: Why So Few?" speaksof a "gender gap in jobs" for women Ph.D.s inthe science and engineering work force. With theexception of engineering, male scientists are ableto realize their plans for either employment orpostdoctoral fellowships sooner than women. Inacademia, men are "far more likely than womento be hired to tenure-track positions, to be pro-moted to tens' - Id to achieve full professor-ships." Women, on the other hand, "hold assis-tant professorships and nonfaculty positions morethan twice as often as men. In industrial research,women Ph.D.s are underrepresented by about 50percent." Those who, do obtain employment areonly about half as likely as men to advance tomanagement positions."

The salary differential between women and menin comparable scientific positions is quite pro-nounced. Data from the National Science Foun-dation (NSF) show that women earn less than menin almost every field of science, in every employ-

"Berryman, 1985, op. cit , p. 7"Lilli S. Hornig, "Women in Science and Engineering: Why So

Few?" Technology Review, November-December 1984, p. 40.

119

ment sector, and at every level of experience."In 1982 the median annual salary of women scien-tists and engineers was 75 percent of that of theirmale counterparts: $26,300 v. $35,000 The per-centage was highest among the comp- te: special-ists at 86 percent, and lowest among the life scien-tists, social scientists, and physical scientists at 74percent. (See table 5-1.) Employed female scien-tists and engineers with less than 5 years' experi-ence earned on .he average 90 percent as muchas their male counterparts; those with 31 to 35years experience earned less than 78 percent.

Aline Quester, in an exhaustive study of "TheUtilization of Men and Women in Science andEngineering Occupations: Task and Earning Com-parability" finds that:"

Male scientists and engineers earn substantiallymore than female scientists and engineers. While

"The 1982 Postcensal Survey of Scientists and Engineers, NSF84-330 (Washington, DC: National Science Foundation, 1984), ta-bles 8-32 and 8-33, pp. 144-151; and Science and Engineering Per-sonnel: A National Overview, op. cit., table 13-17. pp. 126-129. Thesetwo publiritions are the sources for all the numbers in this para-graph. They are not quite consistent with one another.

"Aline 0. Quester, Utilization of Men and Women in Scienceand Engineering Occupations: Task and Earning Comparability(Alexandria, VA: The Public Research Institute, July 1984).

one-fourth to one-third of the male earningspremium is accounted for by differences in in-come-producing characteristics between malesand females (primarily different subfield concen-trations), the other portion of the differential isunexplai:A; the men simply earn more than thewomen.

No observable variables have yet been isolatedwhich would account for the systematic earningsdifferential. If no such variables emerge in the faceof repeated investigation, the presumption growsstronger that the earnings differential rests on cov-ert discrimination.

Unemployment

In 1974, NSF reported unemployment rates[among all scientists and engineers] of 1.6 percentfor men and 4.1 percent fvi women. By 1982, un-employment rates had climbed to 1.9 percent formen and 4.5 percent for women. Unemploymentrates for doctoral men and women scientists an.:engineers were 0.9 and 3.9 percent, respectively,in 1973, and 0.8 and 2.6 percent in 1983.

NSF found the smallest unemployment ratedifferential between women and men among com-puter specialists, while the greatest difference was

Table 5.1.- Average Annual Salaries of Scientists and Engineers by Field and Sex!RacelEthnic Group, 1982

Field Total Men Women White Black AsianNative

American HispanicTotal, all fields $34,000 $35,000 $26,300 $34,100 $29,900 $34,200 $34,000 $31,400Total scientists 31,700 33,400 25,800 31,800 28,500 32,400 32,600 27,600Physical scientists 34,700 35,500 26,400 34,900 30,100 32,500 42,500 33,600

Chemists 33.600 34,600 25,500 33,900 29,500 30,400 42,300 29,800Physicists/astronomers 37,900 38,100 32,600 37,900 34,600 40,500 43,500 40,500Other physical scientists 35,000 35,700 26,300 34,900 33,400 37,100 42,100 39,800

Mathematical scientists 34,800 37,500 29,100 35,000 31,600 34,500 31,200 25,400Mathematicians 35/10 37,700 29,500 35,600 31,800 36,200 31,200 30,060Statisticians 32,.....) 36,700 28,100 33,000 30,900 28,600 (1) 17,200

Computer specialists 32,200 33,500 28,800 32,300 31,100 32,000 33,000 30,600Environmental scientists 36,800 38,000 29,900 36,700 30,7W 37,200 46,600 38,500

Earth scientists 37,600 39,000 30,300 37,500 31,200 38,100 42,200 39,800Oceanographers 34,600 36,500 22,300 33,400 28,200 30,000 56,400 22,400Atmospheric scientists 32,700 33,100 28,500 32,600 29,400 33,600 (1) 31,400

Life scientists 28,900 30,400 22,500 29,000 27,700 28,100 30,800 25,600Biological scientists 28,200 29,500 22,500 28,300 28,000 27,400 25,800 24,100Agricultural scientists 27,500 28,800 17,900 27,400 26,300 28,100 35,700 27,600Medical scientists 38,900 42,600 28,200 39,300 27,100 32,000 34,500 30,700

Psychologists 28,800 31,700 23,900 29,000 25,90P 28,400 23,300 20,400Social scientists 30,600 33,000 24,300 30,700 28,400 34,300 29,000 24,100

Economists 34,700 35,900 29,600 34,700 31,100 37,200 28,700 31,000Sociologists/anthropologists 24,900 27,000 21,600 24,900 23,800 26,700 28,500 18,100

Total engineers $35,800 $36,000 $29,000 $35,900 $31,700 $35,100 $35,000 $33,700SOURCE National Science Foundation, Science and Engineering Personnel' A National Overview, NSF 85-302, p 138

L. ID

120

noted among social scientists. After controllingfor field, the unemployment rate for women re-mained twice that for men. For recent (1989 and1981) science and engineering graduates at thebachelor's level, 7.7 percent of the women and5.1 percent of the men were unemployed. Amongrecent master's degree graduates, 7.3 percent ofthe women and 2.3 percent of the men were alsounemployed."

Underemployment

The term "underemployment" is used by NSFto describe the combined effect of involuntary em-ployment outside of science and engineering andinvoluntary part-time employment where full-time employment is sought. The "underemploy-ment" rate for women scientists and engineers in1982 was 5 percent; for men it was 1 percent. Partof this difference was due to the greater concen-tration of men in engineering, where full-time em-ployment is more the rule. But when only scien-tists are compared, women are still twice as likelyas men to be "underemployed." NSF reports thatunderemployment rates for women are higher inevery field of science except for computer spe-cialists, where the rates are essentially equal. Thisis true also at the doctoral level, where underem-ployment rates for women are above those formen in all major fields of science and engineer-ing."

Rank and Tenure

According to Betty Vetter:2°

Among all academically employed doctoralscientists and emtineers in 1983, 65.6 percent ofthe men, but only 39.2 percent of the women,were tenured. An additional 14 percent of menand 21 percent of women were on the tenuretrack, while 8.4 percent of men and 19.9 percentof women were neither tenured nor in tenure-track positions. . . .

The National Research Council (NRC), in 1981,reported on the results of a survey of Career Out-comes in a Matched Sample of Men and Women

"Women and Minorities in Science and Engineering (Washing-ton, DC: National Science Foundation, January 1984), pp. 18-21

"Ibid."Betty M. Vetter, 'Women in Science and Engineering," type-

script, p. 10.

Ph.Ds. It found that for men and women withdegrees in the same field, in the same year, fromequally prestigious universities, significant gen-der differences could be found in employment,rank and promotion, and salary. Specifically,NRC found that:2'

Among the academically employed Ph.D.s whowere surveyed 20 or more years past the doc-torate, 87 percent of the men were full professorscompared with 64 percent of the women.

For a given pair of one woman and one manwith matched characteristics110-19 years past thePh.D.], the man is 50 percent mere likely thanthe woman to have been promoted to full pro-fessor.

Among 1070-1974 Ph.D.s one-third of thewomen, but one-half of the men held seniorfaculty posts. In every field, the distribution byrank was less favorable for women than men,based on their greater concentration among as-sistant professors and nonfaculty appointees.

Female salaries at major research universitiesare significantly below the estimated salaries formen with similar characteristics.

Salary differences between young male and fe-male Ph.D.s in academe still exist, even after con-trolling for type and quality of doctoral training.

Lilli Hornig reports that only 79 out of approx-imately 4,200 faculty positions in the 171 Ph.D.-granting physics departments in the United Statesare held by women. Women hold only 188 of the4,400 faculty positions in chemistry departmentsthat grant the doctorate.22 This situation exists de-spite the fact that, according to NRC, there aremore than 3,600 women doctoral chemists in theU.S. labor force.23 Vivian Gornick, in her bookon Women in Science likens the situation ofwomen in chemistry to that of "Jews in CzaristRussia."24 She reports the following statementfrom an anonymous woman chemist at a "greatresearch university":25

"Nancy C. Ahern and Elizabeth L Scott, Career Outcomes ina Matched Sample of Men and Women Ph D.sAn Analytical Re-port (Washington, DC: National Academy Press, 1981), pp. xvii,xviii.

"Hornig, op. cit., p. 41."Betty D. Maxfield and Mary Belisle, Science, Engineering, and

Humanities Doctorates in the United States. 1983 Profile (Wash-ington, DC: National Academy Press, 1985), p. 28.

"Vivian Gornick, Women in Science (New York: Simon &Schuster, 1981), p. 98.

"Ibid , pp. 102-103.

1 -.1..., '-.

121

The chemistry department here doesn't adver-tise. It's illegal now, but they still do it that way.Somehow they consider it a "shame" to advertise.They write to their friends. And of course theirfriends are men who have only male graduate stu-dents. But even so, some awfully good youngwomen get through the system and come up herefor interviews. It's always the same. They lookat these excellent young women and they say,"She's very good but she lacks seasoning. Let hergo off somewhere else for the year and then wellconsider her again." Of the young men just likeher they say, 'We'd better grab him before some-one else does."

Implications for Women's Participationin Science and Engineering

Of the many factors that reduce the participa-tion of women in science, and engineering educa-tion and employment, the discriminatory prac-tices discussed in the preceding section are perceivedby many to be the most serious impediments tothe goal of equality of opportunity. Those prac-tices are thought to violate the equity principlemost directly, because they affect people whohave established, by virtue of obtaining an ad-vanced degree, the right to pursue a scientific ca-reer based solely on the quality of their work.Their effect on women who have made the longand arduous investment in training for a scien-tific career can be devastating. Vivian Gomick de-scribes discrimination against women in scienceas:

. . . the kind of experience that becomes lodgedin the psyche: both the individual one and the col-lective one. It may go unrecorded in the intellectbut it is being registered in the nerve and in thespirit. It means sustaining a faint but continuoushumiliation that, like low-grade infection, iscumulative in its power and disintegrating in itsulthnate effect (emphasis added].

The differential treatment of women in thescience and engineering work force is believed tohave a significant effect on female students in theeducational "pipeline." A woman student in aphysics or chemistry department where no womenhave achieved tenure, or, perhaps, even beenhired to a tenure-track position, is not likely toform a positive picture of her likely future em-ployment prospects. Nor will she experience the

mud., p. 74.

kind of role model which the literature on equal-ity of opportunity suggests is desirable to assista young woman in identifying herself with her fu-ture profession. These two factors will consider-ably decrease the motivation for such a studentto make the sacrifices required to stay in gradu-ate school and complete her Ph.D.

Finally, the discrimination, higher attritionrates, greater unemployment, and underemploy-ment experienced by women as compared to menin science and engineering are seen by many tobe a serious waste of human resources that havebeen cultivated and prepared at considerable ex-pense to the individual and the Nation.

The two avenues for dealing with this problemappear to be strict enforcement of existing affirm-ative action laws, and leadership from within thescientific community. As Li Ili Hornig writes:"

Despite the widespread nonenforcement of"affirmative action," laws against explicit biashave opened up much broader access to educa-tion and careers for women. Universities dis-courage most of the more obvious forms of dis-crimination against woman and point with prideto equal access and success for women students. . . . The most effective way to deal with (lessexplicit disparities] is probably not by external in-tervention but through the leadership of adminis-trators and senior faculty. MIT took this ap-proach more. than a decade ago and has hadconsiderable success in recruiting women as bothstudents and. faculty, even in fields that havetraditionally "had no women."

The National Academy of Engineering, in itsreport on Engineering Education and Practice inthe United States," has taken such a leadershiprole. It finds "anecdotal" evidence that "femaleengineering professors are not obtaining tenureat the same rate as their male counterparts"iand"a perception of discrimination against femalefaculty members in assignment of teaching respon-sibilities and in selection for research teams." Itrecommends that college administrators "make acandid assessment of the negative aspects ofcampus life for women faculty members" and,where these are found, "take firm steps to elimi-nate them."

"Hornig, op. dt, p. 41."National Research Council, Engineering Education and Practice

in the United StatesFoundations of Our Techno-Econank Future(Washington, DC: National Academy Press, 1985), p. 94.

122

MINORITIES

Minorities represented 9.7 percent of the scienceand engineering work force in 1982, up fromabout 5.5 percent in 1976, but substantially lessthan their 18.0 percent representation in the gen-eral working population. Blacks constituted 2.6percent of the Nation's scientists and engineers,as compared to 10.4 percent of the general laborforce. Hispanics represented 2.2 percent of thescientific and engineering work force, ; opposedto 5.5 percent of the total labor force. On theother side of the coin, Asian-Americans' 4.5-percent share of the scientific work force wasnearly triple their 1.6-percent share of the work-ing population in the United States."

In the educational "pipeline," minorities' dif-ferential experience in science and engineeringfrom that of white males is also quite dramatic.As table 5-2" shows, blacks, Hispanics, andAmerican Indians receive degrees in quantitativefields at less than half the rate of whites, whilethe rates for Asian-Americans are more than dou-ble that c c the white population. The numbers inthe table represent the ratio of the percent of quan-titative degrees awarded to the particular groupto its percentage representation in the age-relevantpopulation. For example, blacks received 4.1 per-cent of the quantitative B.A.s in 1978-79, but were

"Vetter and Babco, op. cit., p. 96. See also Statistical Abstractof the United States, 1985 (Washington, DC: U.S. Department ofCommerce, Bureau of the Census, 1905), pp. 391-392.

"Berryman, 1903, op. cit., p. 21.

Table 5.2.-1978179 Representation In QuantitativelyBased Fields' Relative to Representation inAge-Relevant Population by Degree Level

and Racial and Ethnic Group

Racial and ethnic group

Degree level

B.A.Professional

M.A. Ph.D. degreesWhites 1.13 1.12 1.12 1.12Blacks 0.32 0.21 0.18 0.35Hispanics 0.55 0.29 0.21 0.47American Indians 0.43 0.50 0.33 0.50Asian-Americans 1.93 2.79 2.71 1.58klusrultatIvety based fields for the BA , MA, and Ph 0 are defined to includethe phyeidal scin.a.s, mathemettcs, computer sciences. biological sciences,angineeting, and economics. For professional degrees the fields 1r biologiaft c phystmay bend and defined to Include medicine, dentistry, optometryaMsopathy, podiatry, veterinary medicine, and Mummy

SOURCE. Sus Berrynten, Who Will Do Selena (New York The Rockefeller Foundation, 1013), p. 21.

12.9 percent of the age 22 population, leading toa ratio of 0.32 in the table.

The very low ratios for blacks, Hispanics, andAmerican Indians are, in fact, the product of twofactors: the tendency of these groups to receivehigher education degrees at far lower rates thanwhites, and their tendency, as well, to major innonquantitative fields. These two factors are dis-played in tables 5-3 and 5-4.31 Blacks, Hispanics,and American Indians are 50 to 62 percent aslikely as whites to obtain a baccalaureate, and 30to 66 percent as likely to receive a Ph.D. Amongthose who do obtain the two degrees, the threeminority groups under discussion are 62 to 88 per-cent as likely as whites to have majored in a quan-titative field at the undergraduate level, and 39

"Ibid., pp. 111 and 20.

Table 54-1978/79 Representation Relative toRepresentation In the AgeRelevant Population by

Degree Level and Racial and Ethnic Group

Racial andethnic group

Degree level

Associatedegree B.A. M.A.

ProfessionalPh.D. degree

Whites 1.04 1.11 1.10 1.11 1.14Blacks 0.70 0.51 0.58 0.41 0.35Hisoanics 0.88 0.62 0.38 0.31 0.45American Indians 0.98 0.57 0.88 0.68 0.50Asianmericans 1.27 1 13 1.05 1.33 095SOURCE Sus Berryman, Who Will Do S& Woo (New York The Rockefeller Foun-

dation, 1964 p 21

Table 5-4.-1978179 Representation in QuantitativelyBased Fields' Relative to Representation in

Total Degrees by Degree Level andRacial and Ethnic Group

Racial and ethnic group

Degree level

B.A.Professional

M.A. Ph.D. degreesWhites 1.02 1.02 1.01 0.99Blacks 0.62 0.36 0.39 0.96Hispanics 0.88 0.80 0.89 1.04American Indians 0.75 0.75 0.50 1.00AsianAmericans 1.71 2.65 2.04 1.67lauantitatively based fluids for the BA,MA, and Ph D are defined to includethe physical KWIC'S, mathematics, computer schwas, biological sciences,engineering, and economics For pro'ossional degrees the fields are bloiogi-cally or physically based and defined to include medicine, dentistry, optometry,osteopathy, podiatry, veterinary medicine, and pharmacy

SOURCE Sus Berryman, Who Will Do Science (New York The Rockefeller Foun-dation, 191)3), p 21

123

to 69 percent as likely t3 have majored in a quan-titative field on the graduate level. The ratios intable 5-2 are the product of the ratios in tables5-3 and 5-4 for each racial and ethnic group andeach degree level.

Asian-Americans are "out equal to whites intheir likelihood of obi.. . higher educationdegrees. However, they are more than twice aslikely to select quantitative majors in college andgraduate school. This picture for Asian - Americansmay, however, be deceiving. Robert Suzuki claimsthat 85 to 90 percent of the Asian American scien-tists and engineers are non-U.S. citizens, ornaturalized citizens who immigrated to this coun-try to pursue their college education. This indi-cates, according to Dr. Suzuki, that:32

American-born Asian/Pacific Americans ofsecond, third and even fourth generation whobear the legacy of 130 years of racial oppressionand who generally trace their ancestry to poor im-migrant peasants are probably not over-repre-sented in science and engineering and, indeed theymay still be underrepresented, although I knowof no definitive studies on this subject.

On the other hand, most of the foreign-bornAsian/Pacific Americans in science and engineer-ing come from the more affluent classes in theircountries of origin and represent perhaps the topone-hundredth of 1 percent of their country's pop-ulations. Consequently, these persons have notsuffered the historical discrimination experiencedby their American-born counterparts. Moreover,they generally represent an elite class, the creamof the cream, who --e likely to do well even asimmigrants.

Dr. Suzuki is undoubtedly overstating the case.However, because the high participation rateamong Asian-Americans in science and engineer-ing education and employment does not consti-tute a problem in equality of opportunity, we willuse the term "minorities" to refer exclusively toblacks, Hispanics, and Native Americans in thischapter. Asian-Americans will be discussed sep-arately, where appropriate.

Due to limitations of time and space, differencesbetween blacks, Hispanics, and American Indians,

"U.S. Congress, House Committee on Science and Technology,Subcommittee on Science, Symposium on Minorities and Womenin Science and Technology (Washington, DC: U.S. GovernmentPrinting Office, July 1982), p. 12.

between men and women in each minority groupand between the different Hispanic subgroups can-not be discussed in this technical memorandum.The omission of a discussion of these differencesis not meant to imply that they are insignificant.

The quality of academic preparedness in sec-ondary school is cited by many experts as thegreatest factor affecting minorities' academic per-formance and baccalaureate attainment in college.The greater attrtion levels in the sciences of mi-nority groups correlates very highly with meas-ures of academic preparedness, such as highschool grades, aptitude test scores, quality ofstudy habits, rigor of the high school curriculum,and perceived need for tutoring. Of those factors,Alexander Astin found that grade average andclass rank were more important predictors of un-dergraduate grades and persistence than werestandardized test scores.33

Unlike some other disciplines, it is essential tobegin the science course sequence at an early stagein the high school curriculum. Fields such aschemistry, physics, and engineering require ex-tensive preparatory coursework. Students in pri-vate high schools who have greater access to col-lege preparatory curicula, including advancedmathematics and science courses, than do studentsin public high schools, tend to choose mathe-matics and science majors in larger proportions.The poor quality of mathematics and science cur-ricula in many inner-city high schools has beenfound to be a contributing factor to the low rateof selection of science and mathematics majorsamong minorities. It has been found that a higherpercentage of black students from predominantlywhite high schools choose mathematics-basedmajors than blacks from predominantly blackhigh schools. In a study of 474 juniors andseniors at Wayne State University, where or s-fifthof the students are black, it was found that:"

Over 70 percent of the white science majors felttheir high schcol training was adequate while less

"Alexander Vi . Astin, Minorities in American Higher Education:Recent Trends, Current Prospects & Recommendations (San Fran-cisco, CA: Jouey -Bau, Inc., 1982), p. 180.

"Thomas, op. cit., p. 8."Maureen A. Sie, et al., "Minority Groups and Science Cat.-oxs,"

Integrate Education, vol. 16, May-June 1978, pp. 44-45; quote* 'IIMeyer Weinberg, The Search for Quality Integrated Education: Pol-icy and Research on Minority Students in School and College (West-port, CT: Greenwood Press, 1983), p. 288.

-130

124

than 50 percent of the black science majors did.Sie and her colleagues found that a number ofblack students had taken advantage of special op-portunities in secondary school: Of the twenty-four black science majors, thirteen went to CassTechnical High School where science is em-phasized.

Because of the high-level of preparation re-quired for the sciences many minorities opt forother fields of study. Astin's study revealed thatwith the possible exception of Puerto Ricans, dur-ing their senior year in high school minority stu-dents already show a strong preference for an edu-cation major and a tendency to avoid majors inthe physical sciences and mathematics and in engi-neering. This was attributed in part to the stu-dents relatively poor academic preparation at thesecondary school level." Sue Chipman and Veron-ica Thomas report that "a survey of high schoolstudents who were seniors in 1980 indicated thatblack, Hispanic, and Native American studentswere only about half as likely to have taken ad-vanced math courses as white students, whereasAsian-American students were about twice aslikely to have done so.""

The educational level of the minority student'sparents plays an important role in determiningwhether the student will be enrolled in an engi-neering or science curriculum. Students whoseparents have obtained college or graduate degreesare more often enrolled in quantitative majorsthan are students who have less well-educated par-ents. Sue Berryman found that "being second gen-eration college not ordy increases but also equal-izes, the choices of quhntitative majors acrosswhite, black, American Indian, Chicano andPuerto Rican college freshmen."" (Asian-Ameri-can students select quantitative majors at muchhigher rates than any other group whether theirparents have a college education or not.) College-educated parents apparently tend to assume theirchildren will also atterd college, and therefore en-courage them to enroll in the required prepara-

"Astin, op. at., pp. 73-74."Susan F. Chipman and Veronica G. Thomas, 'The Participa-

tion of Women and Minorities in Mathematical, Scientific, and Tech-nical Fields," commissioned by the Committee on Research in Math-ematics, Science, and Technology Education of the Nati mal ResearchCouncil Commission on Behavioral and Social Scie ices and Edu-cation, September 1984, p. 49.

"Berryman, 1985, op. cit., p. 17.

tory courses. College-educated parents also ap-pear to be better informed about the importanceof pre-college training, and expose their childrento a greater variety of career options.

The financial resources available to the minor-ity college student play an important role in de-termining academic success and attrition rates.Minority students typically experience difficultyin financing undergraduate study. They must relymore on scholarship, work-study, and loan pro-grams in contrast to nonminority students, whoreceive greater family support. In 1975, black andHispanic college-bound high school seniors esti-mated that their parents would contribute about$200 a year toward college expenses, while themedian figure for whites was over $1,100. Thatsame year minority students comprised one-thirdof the persons assisted through the major U.S. Of-fice of Education aid programs. Upon graduationfrom college, immediate emplcyment opportuni-ties may appear more rewarding than advancedstudy in view of the prospect of further financialdifficulties, the academic risk of graduate study(about half of all doctoral candidates fail to com-plete Ph.D. degrees), and labor market uncertain-ties.3" Associated with these financial difficultiesarc the problems of the working student in gen-eral. If a student must hold down a full-time jobwhile in college he or she is less likely to com-plete his or her baccalaureate."

Factors such as poor academic preparednessand inadequate financial support provide a sur-face-level explanation of why minorities tend toparticipate at lower rates in higher education ingeneral, and science and engineering in particu-lar. Underlying these factors are the deeper issuesof culture and social class. Sue Berryman describesthe importance of these factors as follows:"

Racial and ethnic differences in mat' maticalachievement that we observe at grade 9 appear

"National Board on Graduate Education, Minority Group Par-ticipation in Graduate Education (Washington, DC: National Acad-emy of Sciences, 1976), p. 8

'°Astin, op. cit., p. 109."Berryman, 1985, op. cit., pp. 14-17, the studies cited by Berry-

man to support her conclusions are the following:J. D. Coleman et al., Equality of Educational Opportunity(Washington, DC: U.S. Department of Health, Education, andWelfare, 1966).R.H. Dave, 'The identification and Measurement of Environ-mental Pro, ?SS Afariables That Are Related to Educational

0l

125

at grade 1, [with] blacks, Chicanos, and PuertoRicans starting school wiih mean scores on ver-bal and non-verbal tests of achievement below thenational white average.

Two momentous factors contribute to the rela-tionship between ethnicity and mathematical per-formance at each educational stage: culture andsocial class. Both affect family behavior patternswhich in turn powerfully affect children's schoolperformances.

A study of verbal, reasoning, numeric and spa-tial achievements among Puerto Rican, Jewish,Chinese, and black children at grade 1 shows clearracial and ethnic differences in the patterns ofthese abilities and subsequent studies suggest thatethnic differences in ability patterns at grade 1persist through elementary and secondary school.More important, although social class has impor-tant effects on the level of abilities of each group,it does not alter the basic pattern of abilities asso-ciated with each group.

At the same time, the study shows that middle-class children from the various ethnic groups re-semble each other to a greater extent than scoresof the lower-class children from the differentgroups . . . . Social class has a particularly pro-found effect on the performance of black children,lower class status depressing performance morefor these children than for children from the lowerclasses of other ethnic groups.

Social class seems to be a proxy for family char-acteristics that affect school achievement. For ex-ample, an American study showed that charac-teristics such as the family's press for achievement,language models in the home, academic guidance

Achievement," University of Chicago, unpublished doctoral dis-sertation, 1963.R.L. Flaugher, Project Access Research Report No. 22Patternsof Test Performance by High School Students of Four EthnicIdentities, Research Bulletin RB-71-25 (Princeton, NJ: Educa-tional Testing Service, 1971).C. Jencks, et al., InequalityA Reassessment of the Effect ofFamily and Schooling in America (New York: Basic Books,1972).G.S. Lesser, "Cultural Differences in Learning and ThinkingStyles," Individuality in Learning, S. Messick (ed.) (San Fran-cisco, CA: Jossey-Bass, Inc., 1976).G.S. Lesser, et al., Mental Abilities of Children From Differ-ent Social-Class and Cultural Groups, Monographs of Societyfor Research in Child Development, Serial No. 102, vol. 30,No. 4, 1965.K. Marjoribanks, Ethnic Families and Children's Achievements(Sydney: George Allen & Unwin, 1979).K. Marjoribanlcs, Families and Their Learning Environments(London: Routledge and Kegan Paul, 1979).

provided by the home, indoor and outdoor activ-ities of the family, intellectuality in the homeas represented by the nature and quality of toys,games and hobbies available to the child, andwork habits in the family together correlated at0.80 with children's achievement scores.

An analysis of 1972 data on blacks' choice ofand persistence in a science major found that fam-ily socioeconomic status affects blacks' choice ofa science major. Higher family socioeconomic sta-tus increased the rate of choosing science majors,the effect operating by increasing the mother'seducational aspirations for the student and thestudent's high school mathematical achievement.When white and black students were equated onthe intervening variables, blacks had a higherprobability of choosing a science major thanwhites [emphasis in the original).

Sue Chipman and Veronica Thomas find thatlower educational and career aspirations asso-ciated with lower socio-economic status may un-dermine minority students perception of the util-ity of mathematics . . . ." She adds that it is "quiteponible that minority students, again because oftheir socio-economic status, have still less knowl-edge of the relationship between mathematics andparticular occupational goals than do students ingeneral."42

Betty Vetter reports on a study carried out bythe National Opinion Research Center in 1980 forthe Departments of Defense and Labor." It iden-tified a nationally representative sample of nearly12,000 16- to 23-year old men and women, andadministered to this group the Armed ForcesQualification Test (AFQT), a general measure oftrainability and enlistment eligibilizy for the armedforces. The test showed that youths from highersocioeconomic groups scored higher than thosefrom lower socioeconomic groups; that whiteyouths generally did better than black or Hispanicyouths; but the strongest single predictor of boththe AFQT score and reading ability was themother's educational level. Later analyses sug-gested that the measured correlation of mother'seducation with test performance approximated the

"Chipman and Thomas, op. cit., p. 50."Betty M. Vetter, -The Emerging DemographicsEffect on Na-

tional Policy in Education and on a Changing Workforce." contractorreport prepared for the U.S. Commas, Office of Technology Amen-ment, May-June 1985, p. 50.

1 " 0

126

combined measured correlation of the four vari-ables usually used to determine socioeconomicstatus: mother's education, father's education,average family income, and father's occupationalstatus.

The differences were substantial. Youth whosemothers had completed eighth grade or less scoredin the 29th percentile on the tests. Those whosemothers had completed high school hadan aver-age percentile score of 54. Those whose motherswere college graduates or more averaged 71.

The effect of socioeconomic status or class onminorities' persistence in the science and engineer-ing educationai. pipeline is a cause for optimismamong some, pessimism among others. Sue Ber-ryman paints the optimistic picture:

In the short run, specially designed interven-tions can increase minority shares of quantitativedegrees by targeting those who have the capaci-ties to respond to these interventions, but who,in their absence, would probably not pursue aquantitative training program and career. In thelong run, the trends favor increased minorityrepresentation in the quantitative fields. Thegrowing number of second generation black andHispanic college students will be an important fac-tor in increasing the representation of these groupsin the nation's scientific and engineering laborforce.

Betty Vetter makes the more pessimistic case.She points out that 44 percent of all black house-holds and 23 percent of Hispanic households areheaded by single women. Nearly 72 percent of allblack families and 46 percent of Hispanic familieswith incomes below the poverty level were main-tained by single women. Seventy percent of allblack children are being brought up in poverty.She concludes that:"

We can anticipate an increasing school drop-out rate among Hispanic and black children,growing up with young, single mothers who, be-cause they have too little education themselves,are unlikely to be able to provide the incentivesthat may be required to keep their children inschool and learning . . . . If present trends con-tinue, we can anticipate that more of [the babyboom echo group] will drop out of school or out

"Berryman, 1983, op. cit., p. 4."Vetter, 1985, op. cit., p 7.

of math and science classes at earlier stages, andfar fewer of them will obtain the educational prep-aration required for professional participation inquantitative fields.

In reality, both pictures may be true. An in-creasing r.umber of minority students with mi Idleclass, college-educated parents, will undoubtedlyenroll in higher education and major in scienceand engineering. However, an even larger num-ber of black and Hispanic students from poor,single-parent households will find a college edu-cation and a science or engineering career difficultto achieve due to poor academic preparedness.

Differential Treatmentin the Work Force

As was the case for women, minorities have asomewhat different experience from that of whitesin the science and engineering work force. RecentNSF data" indicate that unemployment ratesamong black and Asian-American scientists andengineers were significantly higher than those ofwhites: 4.6 and 3.3 percent, respectively, versus2.1 percent for whites. The unemployment ratefor Hispanic scientists and engineers, by contrast,was about the same as that of whites, while Na-tive American scientists and engineers had sub-stantially lower unemployment rates. Black,Hispanic and Native American scientists andengineers were somewhat less likely than whitesto be employed in science and engineering (81 to83 percent versus 87 percent). Asian-Americanscientists and engineers were somewhat morelikely (90 percent) to be so employed.

Blacks and Hispanics reported significantlylower salaries than whites in science and engineer-ing. The average salary for whites in all scienceand engineering fields was $34,200; that for blacks$30,100; and that for Hispanics $31,500. The sal-ary differentials varied from lows of $500 betweenwhite and Hispanic environmental scientists and$900 between white and black computer special-ists, to highs of $9,800 between white and His-panic mathematical scientists, and $7,000 betweenwhite and black environmental scientists. Asianand Native American scientists and engineers re-

"Women and Minorities in Science and Engineering, op. cit., pp.viii, ix, 20-24.

127

ported salaries almost identical to those of theirwhite counterparts

Despite these differences in salaries and unem-ployment rates, analysts consulted by OTA did

not report strong evidence for discrimination althe work force against minority scientists andengineers.

EFFECTIVENESS OF PROGRAMS TO PROMOTE PARTICIPATIONAs shown in the preceding sections, many of

the factors that inhibit participation in science andengineering by women and minorities are relatedto pre-college experience, both academic andnonacademic. These factors include overall pooracademic preparedness (for blacks and other mi-no.ities); lack of exposure to the needed scienceand mathematics sequence (for minorities andwomen); socialization factors that underempha-size the desirability or appropriateness of a scien-tific or engineering career (especially for women);poverty and inadequate financial resources (espe-cially for minorities); and family characteristicsthat affect school achievement. To some degree,it appears that these early factors can be over-come, or at least compesated for, by special pro-grams designed to facilitate access to science andengineering education for disadvantaged culturalgroups.

In 1983 the Gffice of Opportunities in Scienceof the American Association for the Advancementof Science (AAAS) conducted an "assessment ofprograms that facilitate increased access andachievement of females and minorities in K-12mathematics and science education" for NSF.49The assessment was based on a survey question-naire to the directors of more than 400 pre-collegeintervention programs for women and minorities,and site visits to more than 50 exemplary projectsin different regions of the United States. Time andbudget limitations precluded AAAS from carry-ing out formal evaluations of the programs it sur-veyed, but the results of such evaluations, wherethey had been performed independently, were re-quested from program dire-tors in the survey.

"Shirley M. Malcom, et al., Equity and Excellence: CompatibleGoals (Washinston, DC: American Association for the Advance-ment of Science, December 1984).

AAAS found that "the primary feature of suc-cessful programs for minorities and females seemsto be that they involve the students in the 'doing'of science and mathematics and convey a senseof their utility." Such "exemplary programs" are"sensitive to the group or groups they are intendedto serve and address these audiences fundamen-tal needs for academic enrichment and career in-formation:"

Exemplary programs for minorities recognizethe deficiencies in performance many students arelikely to have and stress rigorous academic prep-aration in mathematics, science and communica-tions . . . . Projects for females focus heavily oncareer awarenesson the utility of mathematicsand science to whatever they might want to do.Young woman are encouraged to take all thecourses available to them in high school. They areshown models of science and engineering profes-sionals and students who "are making it" in thesefields.

In general, AAAS found that these programs"have demonstrated that there are no inherentbarriers to the successful participation of womenand minorities in science or mathematics," if thesegroups are provided with "early, excellent, andsustained instruction in these academic areas."AAAS also found that:49

Successful intervention programs are those thatha,-e strong leadership, highly trained and highlycommitted teachers, parent support and involvement, dearly defined goals, adequate resources,follow-up and evaluation. For the positive effectsto be sustained, these programs must eventuallybe institutionalized, that is, made part of theeducational system . . . .

Scientists and engineers from the afectedgroups must be involved in the planning as wellas in the implementation of projects . . . .

"Ibid., pp. vii-viii."Ibid., p. viii.

.. ,....J 4

128

Intervention programs must begin early andmust be long-term in nature; "one-time" or short-term efforts do have a place for motivational, in-formational, supplemental or transitional purposes.

AAAS reported on a number of specific inter-vention programs that could document their suc-cess. An evaluation of the Mathematics, Erginetr-ing, Science Achievement (MESA) Program of theLawrence Hall of Sdence in Berkeley, CA, whichincludes 16 centers, 131 high schools and about3,400 students, found that:"

Of recent MESA graduates, 90 percent have at-tended a college or university and approximately66 percent have pursued a math-based field ofstudy. MESA seniors performed significantlyhigher than college-bound seniors of similar ra-cial/ethnic backgrounds across the nation. MESAseniors at sampled schools did not differ signifi-cantly on SAT performance from the total popu-lation of college-bound seniors . . . . despite thefact that the sampled schools were among thelowest-achieving schools in the state.

The Summer Science Enrichment Program atAtlanta University provides summer instructionin mathematics, science, and communication tohigh school juniors, most of whom are black. All338 of the students who have participated in thisprogram since its inception in 1979 have gone onto college, and 95 percent of them have majoredin a quantitative field. It should be noted that thisprogram selects students with demonstrated in-terest and performance in science and mathe-matics.

The Philadelphia Regional Introduction for Mi-norities to Engineering (PRIME) is a consortiumof more than 34 businesses, 14 government andcivic organizations, and 7 universities and pub-lic schools which has operated in the Philadelphiaarea for more than 9 years. It is a supplementary

p. 56.

program in science and mathematics which be-gins in seventh grade and takes students throughhigh school. Of the more than 820 high schoolseniors who have graduated from this programsince 1977, more than 60 percent have chosencareen in engineering and/or technology. In addi-tion, the number of minority students in the Phil-adelphia area enrolled in academic-track highschool programs has tripled during the years ofoperation of the PRIME project, with one-thirdof those students in the PRIME project.

The Professional Development Program (PDP)of the University of California, Berkeley, is afaculty-sponsored program which recruits sopho-mores from 45 public and private high schools ineight districts to participate in special summer aca-demic programs and Saturda' classes during theschool year. Over 60 percent of the students arewomen, and 75 percent are black or Hispanic. Ofthe 421 students from 60 local schools that havecompleted the program, "90 percent have goneon to college and a substantial number are inquantitative fields."61 The average SAT mathe-matics scores of the 1981-82 PDP senior class was598.

These projects illustrate the general point madeby AAAS that intervention programs at the pre-college level can be effective in increasing thepar-ticipation of women and minorities in science andengineering education. A far more systematic andthorough evaluation would be required to docu-ment the extent to which changes observed canbe reliably attributed to the effects of the programsthemselves. It should be noted that participationin these programs is voluntary, so those affectedby the programs tend to be students who are al-ready somewhat motivated toward science andengineering.

"Ibid., p. 56.

POLICY IMPLICATIONSThe AAAS report on Equity and Excellence:

Compatible Goals, cited above, states that "themagnitude and complexity of the problem" ofequality of access to scientific and engineeringcareen requires "a large and continuing effort that

specifically targets large sectors of our society. . . ." The NSF Authorization Act of 1981, Sec-tion 35(a), required the President to submit toCongress, by January 29, 1982, a report "propos-ing a comprehensive national policy and program,

AO

129

including budgetary and legislative recommenda-tions, for the promotion of equal opportunity forwomen anci minorities in science and technology."That report has neither been prepared nor trans-mitted to Congress. Therefore, there is, at thiswriting, no national policy or program to promote equal opportunity in science and engineer-ing for women and minorities.

The Director of NSF did submit a report toCongress on December 15, 1981, in conformitywith Section 34(b) of the 1981 NSF AuthorizationAct which required "a report proposing a com-prehensive and continuing program at the Foun-dation to promote the full participation of minor-ities in science and technology." However thereport contained neither budgetary nor legislativerecommendations as required by the Act, andcontained little more than restatement of existingpolicies and programs." In fact the report at-tempted to rationalize budget cuts in a numberof programs that were created in the 1970s, in-cluding the Minority Institutions Science Improve-ment Program (MISIP), the Resource Centers forScience and Engineering Program (RCSF), the Stu-dent Science Training (SST) Program, the Oppor-tunities for Women in Science Program (OWS),and the Visiting Professorships for Women Pro-gram (VPW).

In the absence of executive branch leadershipin this area, the AAAS study recommended thatthe following steps be taken by the Federal Gov-ernment:53

Federal support for programs to imp -ve thequality of pre-college education in science, math-ematics, and technology should require thatproposals specifically address themselves to plansfor serving women, minority, and disabled stu-dent populations.

Federally supported programs for teacher train-ing and retraining should require that teachingmethods and career and equity aspects be in-cluded, along with a rigorous focus on improvingcompetence in subject content.

The Federal Government should support dis-semination of models previously shown to be

"National Science Foundation, "Proposals of the National ScienceFoundation to Promote the Full Participation of Minorities andWomen in Science and Engineering," typescript, December 1981.

"Malcom, op. cit., p. 30

effective in improving science and mathematicseducation for women and minorities, includingtechnical assistance on management and evalua-tion systems . . . .

. . . . previously supported programs that hada strong positive educational impact on womenand minorities should be reexamined for possiblereinstitution. Of pariicular interest in this regardare the RCSE and SST programs.

In order to better understand the policy impli-cations of the problems experienced by womenand minorities related to participation in scienceand engineering education, OTA sponsored apanel discussion among experts" in this area onJuly 2, 1985. The findings of the panel are pre-sented below:

Issue 1Keeping Options Open

1. The self-perception of women and rinoritiesof their i- ability to succeed in science andmathematics courses is frequently reinforcedby the system's p.-.rception of their inabilityto do science and engineering.

2. Opportunities shot.ld be provided for thispopulation to experience success in scienceand mathematics courses prior to grade 9 aswell as opportunities for them to perceive thevariety of career options and lifestyles thatare based on these disciplines.

3. The Federal Government should support im-provements in the training of junior highschool science and mathematics teachers, thedevelopment of counseling programs involv-ing teachers and parents and the identifica-tion and funding of model programs whichenhance the self-perception of students.

4. Tests should be developed that are better in-dicators of the potential of women and mi-norities to succeed in science and mathe-matics careers.

Issue 2Reducing Attrition

1. The graduate student pipeline should be en-larged by providing long-term support forpromising minority and women studentswho are satisfactorily progressing toward the

The panel consisted of Lloyd Cooke, Michael Crowley, MariaHardy, Shirley Malcom, Shirley Mc Bay, Denis Paul, Willie Pear-son, Luther Williams, Allan Hoffman, and Jerrier Haddad.

130

Ph D. degree in science and engineering withidentification and support beginning at thejunior year in college.

2. The circumstances which lead to the successof women and minority students in scienceand engineering should be studied, and theknowledge obtained applied to improvingthe retention of i, ms successful students.

3. Opportunities for early experience in re-search should be provided for minority andwomen students beginning at the undergrad-uate level. Existing efforts at the graduatelevel should be strengthened.

4. The Federal Government should disseminateand encourage programs and operating con-ditions which demonstrably facilitate thlretention of women and minorities in scienceand engineering.

5. Federal Government affirmative actionguidelines for recipients of Federal funds

should be extended to protect against the fol-lowing:

sexual harassment of women students, andbias of some foreign professors whose cul-tures hold women in low status.

In addition to the above option. .or increasingthe pool size and reducing attrition, it is probablethat further gains might bz realized if more wereknown about how different minority subgroupsrespond to different options. For example, differ-ences may exist:

for blacks, American Indians, and Hispanics;within the various Hispanic subgroups (e.g.,Chicanos, Puerto Ricans, and Cubans); andfor minority and nonminority women.

This suggests the need for support of studies onthis issue. Further suggestions are provided in ap-pendix B.

w

t ,: A ..-7- I

Appendix A

The Education and Utilization ofBiomedical Research Personnel

Introduction

As a nation, we spend $300 billion per year onhealth care, or 10 percent of the gross national prod-uct (CNP). National attenEon to biomedical .esearchand its potential for health benefit has been consist-ently strong since Wo:ld War II, driven by the hopethat science will provide us with longer and healthierlives.

These early beginnings lead to a relatively high levelof Federal support for both research and training inthe biomedical sciences in comparison to other dis-ciplines. The Government, through the National In-stitutes of Health (NIH), made an early commitmentto the training of biomedical research personnelthrough the establishment of the NIH training and fel-lowship programs. In addition, Congress has ensureda substantial and consistent level of support for bio-medical research through favorable appropriations toNIH, the parent agency for federally funded biomedi-cal research programs. As a result of this two-prongedFederal approach to the biomedical sciences throughsupport of both research and training programo, thebiomedical research field has suffered few of the criti-cal personnel shortages and/ or surpluses found inother scientific disciplines. It is, therefore, worth exam-ink tg the Federal program for training and support-ing biomedical research perso.mel in some depth. Itis an example of a continuous and mostly successfulFederal personnel policy in an important scientificfield.

Basic Biomedical Research Personnel[Ph.D.s]

Training

The higher education training system in basic bio-medical research consists of at least two, and oftenthi ee, stages. Virtually all biomedical scientists obtaina doctorate, although exceptions do exist. In addition,the Institute of Medicine (IOM) estimates that 60 per-cent of Ph.D. recipients in the basic biomedical sciencesgo on to postdoctoral appointments. Postdoctoral ap-pointments are considered by some +o be a holdingpool for the Ph.D. population in -ch of employ-ment. Recently, it has been suggestet at postdoctoraltraining is also an essential educational experience for

the biomedical scientist, and essential to su:cess as anindependent investigator.' Rapidly developing areasof research, it is arc-,ed, require more intensive andext..nsive training.

IOM estimates that c the postdoctoral populationwith doctorates from L.S. universities 30 percent arewomen and 10 percent hold foreign citizenship.' For-eign scientists have comprised approximately one-thirdof the total population of biomedical postdoctoral ap-pointees since 1975.

The typical biomedical researcher spend 4 years inundergraduate work, 5 to 7 years in graduate schoolto receive the Ph.D., al. 2 to 4 years in a postdoc-toral appointment. Figur? A-1 illustrates the doctoraltraining system in the biomedical sciencef..

Figure A-2 shows the distribution of primary sourcesof support for graduate students enrolled in Ph.D.-grai iting biomedical science departments, in 1975 and1981. In addition to Federal sources of support, bio-medical science students rely on teaching assistantshipsand self-support. NIH is the major source of supportfor postdoctoral training in clinical research, supply-ing nearly 90 percent of all training funds.' The Na-tional Science Foundation (NSF) provides for approx-imately 45 predoctoral fellowships per year in thebiomedical sciences.*

It is only in the past 12 years that Federa' programsfor biomedical research training have been centralized.In 1973, the Nixon Administration proposed phasingout all training grant and fellowship programs of NIH;the Alcohol, Drug Abuse, and Mental Health Admin-istration (ADAMFIA); and the Health Resources Ad-ministration (HRA). Congress responded, in 1974, bypassing the National Research ServiceAward (NRSA)Act (Title I, Public Law 93-346) that consolidated pre-vious research training authorities under the PublicHealth 5- lice Act. The NRSA Act is the only exist-ing authority under which NIH supports training forbiomedical and behavioral research careers. Thr!atsto the NRSA r ogram have appeared in every admin-

'Institute of Medicine, Personnel Needs and Training for Biomedi-cal and Behavioral Research: 1943 Report (Washington, DC: Na-tional Academy Press, 1983)

'Ibid., p. 61.'Thomas E. Malone, "Assuring the Supply 3f Clinical Investiga-

tors" (typescript), National Institutes of Health, 1981.'Personal communication with Terence Porter, Science and Engi-

neering Education, National Science Foundation, March 1985.

133

134

Figure A.Doctoral Training System in the Biomedical Sciences

1,200

(Est. sin In 1961: 96,100)

(Est. size in 1961: 80,900)

NOTE Estimates represent the avenge annual number of Individuals following particular pathways during the 197341 period. i4o elltIMMIMI hare been made forimmigration. emigration. or reentry into the [abr force.

SOURCE Ku Report on Personnel Heeds, 1553.

istration budget between 1978 and 1985 only to be'relieved by Congress.

NIH devotes 10 percent of NRSA appropriations tospecial programs in minority access, the Medical Sden-tilts program leading to a combination M.D./Ph.D.degree, and short-term training programs in medicalschools. Predoctoral training accounts for 35percentof the NIH NRSA program. The program's major em-phasis, however, is on postdoctoral training, whichreceives 55 percent of NRSA funds in the belief thatchanging national needs can be most quickly met bya large and well-trained post('-ctoral pod.

Funds are awarded for preductoral and postdoctoralstipends through a system of individual fellowshipsand institutional training grants. Institutional traininggrants must survive peer re Aew in national competi-tion, and require a substantik4 commitment on the partof the institution to share in the maintenance of a train-ing environment. They are limited to those institutionswith the tradition and history of producing high-quality investigators and the financial resources to ac-cept the necessary cost-sharing burden. Institutionaltraining grants provide the program director with anumber of full-time training positions. Individual fel-lowship awards, on the other hand, allow the appli-cants to develop a research training project under the

supervision of a sponsor. The individual fellow tendsto be further Iloug in the training process than the in-stitutional fellow (i.e., at the postdoctoral level). In1985, the average annual stipend for the predoctoralellowship is $o,552; the range for postdoctoral awardsis $16,000 to 21,000 irr year.

By 1971, Nat (raining grants and fellowships sup-ported or assisted 37.3 percent of the Nation's full-timegradua students in the medical sciences and 1 per-cent in tne life sciences.' Between 1961 and 1972, NIHp.,ogrants furnished financial assistance through fel-lowships and institutional training grants to more than30,000 graduate (predoctoral) students in the health-related disciplines. In addition, more than 27,000biomedical scientists received postdoctoral supportth-ouet NIH programs.' Since 1972, individual awardsfog Tredoctoral students have been eliminated and thentur:oors of predoctorals on training grants have beenreduced. Between fiscal years 1969 and 1980, the num-ber of predoctoral students receiving NIH supportdropped by about 50 percent (see figure A-3). Never-theless between 1973 and 1983, more than one out

'Malt t, op. cit.'National Re.earch Council, Office of Scientific and Engineering

Personnel, Survey of Earned Doctorates, 1920411.

l

/1111

:135

Figure A2.Distribution of Primary Sources ofSupport for Graduate Students Enrolled in

Ph.D.Granting Biomedical ScienceDepartments, 1975 and 1981

1975

1981

NIHtrainees'

(12%)

ratietantshig ::g4t/e) , -

SOURCE Natic-a1 Science Founda.ion. Survey of Graduate Science Studentsand Postdo:torals Computerized Data File VVeshingf on DC 1973 83

Figure A3.Number of Predoctoral Trainees andFellows Supported by NIH, Fiscal Years 1967.8010,000

9,000

8,000

7,000

a

Tia

a,

m6,000

0

85,000

4,000

E

z 3,003

2,000

1,000

067 58 89 70 71 72 73 74 75 76 77 78 7880

Fiscal year

SOURCE 10M Report on Career Achievements, 1984

of every three Ph.D. recipients in health-related fieldshad received some NIH training suppert while in grad-uate school.'

Allocation of Resources Between Predoctoraland istdoctoral Training

The tradition of postdoctoral training in the 'plu-m. 4ical sciences is a long one, possibly 9ecause NIHtraining programs are the oldest of their Kind. The bio-medical research system shows great sensitivity to thesupply and demand for ind:viduals in both groups.The amount of support available for postdoctoral v.predectoral training has a definite effect on the sup.

'Institute of Medicine, Report of a Study. The Career Achieve-ments of N1H Predoctoral Trainees and Fellows (Washington, DC:National Academy Press, 1984)

I

136

ply of manpower in the biomedical sciences. Startingin the early 1970s through 1981, the number of post-doctoral appointments increased 9 percent per year.'(See figure A-4.) During this same period, Ph.D pro-duction was fairly constant. Most of the postdoctoralexpansion has occurred in medical schools, which em-ploy 65 percent of the postdoctoral scientists in bio-medical sciences. In fact, most of the growth has beenin the 20 largest universities.'

Steady growth in the postdoctoral population maybe a combined result of an increase in the numbers ofnew Ph.D. recipients and an increase in the length oftime spent in postdoctoral appointments. IOM believesthat the prolongation of apprenticeship has been a ma-

'Institute of Medicine, Personnel Needs (Washington, DC: Na-tional Academy Pass, 1983).

"National Science Founda.:on, Academic Science. R&D Funds1975-1983, Survey . of Science Resource Series (Washington, DC:U.S. GoverLment Printing Office (annual)).

Figure A-4.Estimated Growth in the Numbers ofPh.D. Awards, Academic Postdoctoral Appointments,and First-Year Postdoctoral Apppointments in the

Biomedical Sciences, 1973-819,000

7,000

3,000

1,000

01973 1976 1977

FiliCtli year

OURCE 10M Report on Personnel Needs. 1963

1979 1961

jor factor in the expansion of the postdoctoral poolin the biomedical sciences. 10 The reason for the ex-tension of postdoctoral training is not clear. A 1976survey revealed that more than 40 percent of the post-doctorals had extended their appointments becausethey were unable to find a desirable position." In 1981,the last year for which data was available, the NationalResearch Council estimated that the postdoctoralgroup constituted 18 percent of the full-time equiva-lent bioscience research personnel in Ph.D.-grantinguniversities."

A particularly unique system of assessment andfeedback in the area of biomedical and behavioral re-search personnel needs was established through lawin 1974 when Congress passed the NRSA Act. ThisAct mandated that the National Academy of Sciences:

. . . esiablis (A) the Nation's overall need for biomedi-ca. and behavioral research percemel (B) the subjectareas in which such personnel are needed and the num-ber of such personnel in each such area, and (C) thekinds and extent of training which should be providedsuch personnel."

In 1975, the committee on a study of National Needsfor Biomedical and Behavioral Research Personnel wasformed under the aegis of the Commission on HumanResources, and later transferred to IOM. During itsfirst 9 years, the committee published seven reportsprojecting the near-term demand for research person-nel in biomedical and behavioral research.

The potential influence of this committee is impres-sive, as it is mandated to make recommendations tothe Federal Government on the size of federally spon-sored training programs for biomedical and behavioralscientists. It also makes recommendations for alloca-tion of traineeships between pro doctoral and postdoc-toral students. To a large degree, the agencies havebeen responsive to these recommendations. The dis-crepancy between the level of support recommendedby IOM and the actual funding level set by NIH hasbeen small in recem years. Decisions regarding distri-bution of traineeships and fellowships, however, areleft to the agency.

In recent reports (1981 and 1983), the committee hasrecommended that the agencies change the allocationof training grants from predominantly predoctoraltraining to predominantly postdoctoral. This recom-mendation was made in response to the growth of the

"institute of Medicine, Personnel Needs, op. cit., p. 60."P.E. Coggeshall, et al., "Changing Postdoctoral Career Patterns

for Biomedical Scientists," Science, vol 202, Nov 3, 1978, pp487.493.

',National Research Council, Postdoctoral Appointments and Dis-appointments (Washington, DC: National Academy Press, 1981).

"U.S. Congress, Title I of the National Research Act of 1974 (Pub-lic Law 93-348), Section 473

.7 4 2

137

predoctoral population that occurred in the 1970s. Thehigher levels of Federal support for predoctoral train-ing relative to levels of support for postdoctoral train-ing which occurred in the 1970s contributed, in part,to the current slight surplus of biomedical Ph.D.s.

Harrison Shull, former Chairman of the Commis-sion on Human Resources of the National Academyof Sciences, believes that the agencies should use theirability to shift funds quickly between predoctoral andpostdoctoral support as a means of responding to afluctuating job market. Postdoctoral funds can be in-creased when the supply of new Ph.D.s is high anddemand is low; when supply is insufficient to meet de-mand, funds can be shifted the other way, towardsmore predoctoral support. This "fine-tuning" mecha-nism can work only if the supply of traineeships is con-trolled and steady."

Demographics and Employment Outlooksfor the Biomedical Ph.D.

The number of doctorate degrees awarded annuallyto bilscientists has been relatively constant since 1972.Bioscience baccalaureate degrees have been decliningsince 1976, and bioscience graduate ersollments peakedin 1978. In the long term, the attrition rate from theactive pool of researchers will increase in the 1990s asthe population ages, creating more academic positions.If graduate school enrollments continue to decline asthe size of the college-age population declines, therecould be a small decline of Ph.D. output in the late1980s and 1990s. If this decline does, in fact, occur,there could be a temporary shortage of biomedical re-search personnel in the next decade.

The key factors 4nfluenoing potential shortages orsurpluses in this market are Federal funding for bio-medical research, and graduate and medical school en-rollments. If Congress continues to appropriate asteady level of funding to NIH, then, in a sense, thedemand is set. Should the Federal Government sud-denly increase or decrease its level of research supportin the biomedical sciences, shortages or surpluses couldoccur.

The total biomedical Ph.D. labor force approached70,000 in 1981. Half of these individuals were em-ployed in academic institutions whose population hasbeen growing at a rate of 4 percent per year. Employ-ment in the industrial sector has been growing at a rateof 8.5 percent and self-employment at a rate of 24.7percent (see table L-1 ).

Next to academia, the industrial sector is the sec-ond largest employer of biomedical scientists. Most

"Harrison Shull, "The Ph.D. Employment Cycle -- Damping theSwinp," The National Research Council :n 1978: Current Issuesand Studies (Washington, DC: National Academy of Sdsnces, 1978),pp. 149-163.

biomedical Ph.D.s in industry are employed byical and pharmaceutical manufacturers. Recently, thebiotechnology industry has shown rapid growth, andit is possible that some areas of commercial biotech-nology may experience temporary shortages, as in-vestigators attempt to catch up with the latest devel-opments in a new field. A collaborative survey ofbiotechnology firms conducted by IOM and the Of-fice of Technology Assessment in 1983 revealed thatone-third of the respondents had experienced short-ages in one or more specialties. The three specialtiesmost often cited were bioprocess engineering, recom-binant DNA molecular genetics, and gene synthesis."

Emerging fields will experience temporary shortagesof trained personnel as new research opportunities be-come available. It is the view of some that the post-doctoral pool will be the critical element in ensuringan efficient and timely transition to steady-state sup-ply, for postdoctorals are the most skilled in researchmethods and problem-solving and can most easilyadapt to new concepts and new tools.

Clinical Biomedical ResearchPersonnel 111.D.si

The application of scientific advances to maintaingood health and prevent and treat disease is the goalof clinical, as opposed to basic biomedical research.For the purposes of this report, clinical investigationincludes research on: 1) patients or samples derivedfrom patients as part of a study of the causes, mecha-nisms, diagnosis, tre; nest, prevention, and controlof disease; or 2) research on animals by scientists iden-tifiable as clinical investigators on the basis of otherwork."

In general, clinical investigation requires the partici-pation of an investigator trained in the clinical sciencesand possessing a degree in medicine (M.D., D.V.M.,D.D.M., D.0.). It is usually conducted by physician-investigators working in a clinical department of amedical school or in a clinical division of an institute.In general, clinical research is applied research, in-tended to ameliorate disease in the near term. Tradi-tionally, it is the clinical researcher who applies thescientific and technological skills acquired throughbasic research to the vital tasks of medicine.

Both basic and clinical biomedical research ,:encon-nel are critical to the biomedical research endeavor be-cause of the skills and perspectives they bring to their

"U.S. Congress, Office of Technology Assessment, ConintercWBiotechnology: An International Analysis (Washington, DC: U.S.Government Printing Office, January 1984).

"This definition is the one adopted by the Committee on NationalNeeds for Biomedical and Behavioral Research Personnel of the In-stitute of Medicine, National Academy of Sciences.

'43

138

Table A1.-Current Trends in Supply/Demand Indicators for Biomedical Science Ph.D.s

Fiscal yearGrowth rate

from 1975 tolatest year

Latestannualchange

Average annualchange from 1975 to

latest year1975 1976 1977 1978 1979 1980 1981

1. Sway lark:Mrs (sear erresets):a. RID productionb. Percent of Ph LI s without specific

employment prospects at time ofgraduation

3.515

64%

3.578

63%

3.465

72%

3.516

60%

3 644

52%

3 823

48%

3 838

55%

1 5%

-24%

0 4%

146%

54

-02%c Postdoctoral appointments 5.484 n/a 6 403 n/a 7 419 n/a 8.185 6 9% 5 0%a 450d. B.A. degrees awarded 50.493 52,642 51 783 49,701 47 717 45 106 41.476 -3 2% -8 0% -1,503

1 1anasI halleslen MD letelly):a. N a t i o n a l expenditures for health-

*sled SSD (1972 S. bill )b Biomedical science R&D

expenditures in colleges anduniversities (1972 S. bd1 )

c. NIN research grant expenditures(1972 1 W.)

$3 70

$1 19

SO 897

$3 81

$1 26

SO 944

$3 97

$1 27

$1 00

$4 13

$1 34

$1 06

$4 31

$1 35

$1 17

$4 48

$1 43

$1 20

$4 44

$1 49

$1 19

3 1%

3 8%

4 8%

-0 9%

4 2%

-0 8%

SO 123

$0 05

SO 05d. Ph.D. faculty/student ratio°

S. WV &era0 058 n/a C 056 n/a 0 063 0 067

lest ln/a 2 2% 6 3% 0 002

a. Totalb Academic (troluding

postdoctoral's)

40.698

28 339

n/a

n/a

55 167

30 390

n/a

n/a

62 590

33,687

n/a

n/a

69.076

36.497

5 3%

4 3%

5 1%

4 1%

5.063

1.360c. Business 6.645 n/a 6 918 n/a 8.461 n/a 9.957 7 0 A 8 5% 552d Government 4.522 n/a 4,567 ri.'a 5.075 n/a 5.40' 3 0% 3 "c 147e. HOSgitals/chnics 1,950 n/a 2,296 n/a 2 272 n/a 2.79 6 2% 1 3% 141f. Nonprofit 1,221 n/a 1 544 n/a 1.854 n/a 2 083 9 3% 6 0% 144g SoffilnPloYoti 836 n/a 864 n/a 1 197 n/a 1.860 14 3% 74 7% 171li. Other (including postdoctorates)c 6 667 n/a 8 642 n/a 8,899 n/a 9.644 6 3% 4 1% 496I. Unemployed and seeking 518 n/a 810 n/a 690 n/a 831 n/a 9 7% 52

4. tiondled earellnests:a. *WM* 38.314 39.3e2 39.260 43,787 43,378 42 969 42 117 1 6% -2 0% 634J Medical and dental schools 74,222 77.011 79,279 81,934 84.761 86 502 86.254 2 9% 2 0% 2 339c. Estimated undergraduate°d. Total biomedical graduate and

uodergraNiate enrollment

424,539

537,075

439,946

556.279

425,863

544,402

406.373

532,094

374.869

503.008

386.236

515 707

n/a

n/a

-1 9%

-0 8%

3 0%

2 5%

-7,661

-4,1-1aart,e tees CND we not swam for 19e0 latest annual change represents avt,agerola of acaeweally wawa Ph 0 s to a 4 -year *Wed aver.qe of total graduate and undergraduate enrollments ( where (WSh - 1/6(Si + 251-1 + 251-2 + SI-3)Mss Nebdes FFP.1C iaboasories

by Ow formal (Ai +2/8, +2)C, where it, twimeocal science undergraduate enrollment in year I h +2 okeobe.tal selote tatalaurste depress awarded on year + 2 (extudingIts professions). 111_2 mai baccalaureate owns awarded in year I + 2 C, - total undergraduate degree-creon enrollment in year I (excluding erg prolessonal)SOURCE MOM at ellerepot Perna* Reeds and Training for &omechcai and Behavioral Research 1983 p 52

growth rite from 1979-51

work. However, the training of, supply of, and de-mand for clinical investigators present different issuesthan those of basic biomedical research personnel.

Recruiting and Retaining Physicians forResearch Careers

It is the view of many in the medical profession thatthe clinical investigator is the critical link between thelaboratory bench and bedside. While the basic bio-medical scientist is well versed in the processes and in-tricacies of individual biological subsystems, and mayhave better training in research methods and te-li-cliques, it Is the clinician who sees the whole systemand recognizes the clinical manifestations of the under-lying disease process. Thus, research initiated byphyudan-ipvestigators tends to be more health-related, in general, than that of Ph.D. biomedicalscientists.

In recent years, there has been concern that thecountry is facing an insufficient supply of well-trainedphysidan-researchers, as fewer medical school grad-uates enter research careen and apply for research

funds. This prospect is of concern because of the ex-pense of allowing a prolonged lag time between theacquisition of knowledge and technological skills, andtheir application to medical care.

Currently, NIH supports almost 90 percent of allpostdoctoral training in clinical research. The numberof physicians seeking research training has declined onaverage by more than 6 percent per year since 1975while the number of medical school graduates has risenfrom 12,716 in 1975, to 16,347 in 1985." " Accmd-ing to NIH, 47 percent of 5,680 individuals seekingpostdoctoral research training in 1975 were holders ofclinical degrees. By 1980, only 36 percent of 5,321 in-dividuals seeking postdoctoral training held clinicaldegrees." NIH cites several pos dble re.ons for the

"Institute of Medicine, Personal Neel', op. alt."Anne E. Crowley, et al., 'Undergraduate MAW Educe. -n,"

Journal of the American Medical Awodation, vol. 234, No. 12. Sept.27, 1985, pp. 1563.1572.

'llamas E. M alone, "Report on Recruiting and Retaining Phy-sicians for Research Canvas" (typescript), National Institutes ofHealth, 1983.

144

139

decline in the number of physicians seeking researchtraining. Young physicians may be discouraged fromentering research careers because of the increasing dif-ficulty of staying current in two fields (clinical and re-search). There may be too little exposure to researchin medical school because of the competing demandsof an increasingly overloaded curriculum. Young phy-sicians, incurring a high degree of indebtedness aftermedical school, may not be willing to face the tra-ditionally lower pay of research positions versuspractice.

At the same time that interest in research seems tobe declining among medical students, demand forfaculty in the clinical departments of medicine remainsstrong, possibly because of the increase in professionalfee income as a growing source of revenue. Fishmanand Jolly have noted that the expansion in the num-ber of medical schools over the past 30 years led toa greater than 25-fold increase in the full-time clinicalfaculty at American medical centers." As of 1982,IOM reported that the market opportunities for clini-cal investigators continued to be favorable: Medicalschool faculties are still growing 7..ad providing placesfor young physician - investigators interested in researchcareers."

How-ver, after more than 30 years of continuousgrowth, there appears to be a decline in medical schoolenrollments that could continue as the college-age pop-ulation declines. These declining enrollments could po-tential. reduce the demand for physician-investigatorsat medical schools, the principal employers of NIH-sponsored clinical researchers. Thus, although thereis a recognized and widespread realization that thephysician plays a critical role in clinical investigation,and current demand is fairly high, opportunities forthis population could diminish.

Another factor that affects the demand for physi-cian investigators relates to the sources of revenue formedical schools. The abundance of Federal support forbiomedical research since World War II created a con-tinuous demand for trained biomedical research per-sonnel, and provided a sizable source of income formedical schools with active researchers on their fac-ulties.

Although research funds are intended primarily forthe conduct of research, they also enable medicalschools to expand the size of their Faculties, and sup-port graduate students, research assistants, and fel-lows. Federal research support grew from 11 percentof medical school income in 1947 to nearly 30 percentin 1969' However, in recent years, this proportionhas diminished, as medical schools rely increasinglyon other sources of support. In 1983-84, Federal re-

"A.P. Fishman and H. Paul Jolly, "Ph.D.'s in Clinical Depart-ments," The Physiologist, vol. 24, 1981, pp. 17-21.

"Institute of Medicine, Personal Needs, op. cit."National Academy of Sciences, Institute of Medicine, Costs of

Education in Health Professions (Washington, DC: National Acad-emy Press, 1974).

.l.

search funds comprised only 14.4 percent of medicalschools' restricted revenues." The uncertainty of rely-ing on competitive grants as a source of revenue, com-bined with a larger population in competition forlimited funds has led medical schools to seek morereliable and controllable sources of revenue. Fundsfrom professional fee income have grown at an annualrate of 25 percent between 1968 and 1981." The con-tribution of professional fee income to medical schoolrevenues is now 32 percent of total funds as comparedto 12.2 percent in 1970-71.'5 More recently, researchfunds have become available from nongovernmentalsources. As Jolly and his colleagues note in their 1935report on medical school finances:

The gr.mth of private, nongovernmental funding forhealth research and development is a phenomenon ofrather recent nrigin. It may owe a part of its growthto the investment tax credits legislated in the early 1980sfor increases in research and development by industryand the rapid expansion of research investment by thepharmaceutical industry related to rDNA technology,a part of which involves medical school performers.

Effectiveness of Federal Programs

In 1984, IOM, in consultation with NIH, examinedthe extent to which NIH-supported graduate studentshave been successful in their pursuit of careers in bio-medical research." They reported that former NIHpredoctorals have outperformed individuals receivingno direct NIH support in terms of several measures.More than two-thirds of the N7H-supported groupcompleted their doctorates compared to less than 50percent of those receiving no support. In addition, theNIH group were more likely to have received NIHpostdoctoral support and become involved at laterstages in their careers in NIH-sponsored activities.They were more likely to have applied for and receivedNIH research grants. Last, former NIH trainees havewritten more articles and have received more citationsthan their bi 'medical colleagues who received nosupport.

Establishing cause and effect in any program evalu-ation is difficult. The suc.ess of the NIH-supportedgroup may be due to the assistance they received, butit might also be due to the higher levels of motivationand intelligence which caused them to be selected inthe f'..-st place. If so, there was a selection bias fromthe start and this group might have achieved the vari-ous accomplishments without Federal support. Whether

"H. Paul Jolly, et al., "U.S. Medical School Finances," Journalof the American Medical Association, vol. 254, No. 12, Sept. 27,1985, pp. 1573-1581.

"Institute of Medicine, Personal Needs, op. cit."H. Paul Jolly, et al., "U.S. Medical School Finances," Journal

of the American Medical Association, vol. 252, No. 12, Sept. 26,1984, pp. 1533-1541.

"Jolly, et al., 1985, op. cit., p. 1577."Institute of Medicine, Career Achievements, op. dt., 1984.

45

140

they would have been able to complete the Ph.D. with-out Federal financial assistance is the critical question.It appears that the likelihood of completion has, infact, increased.

Summary

Consistent Federal support for biomedical researchin terms of funding of research and training has pro-duced both a substantial supply and demand for in-vestigators. This has had a largely positive effect. Ithas created a steady-state system in which there havebeen no major surpluses cr shortages of biomedicalpersonnel. However, it is a large system that wouldbe difficult to reduce in significant ways without ma-jor dislocatims. Unlike many fields of science, thereis an institutional feedback mechanism between NIHand IOM that ensures a periodic assessment of the sta-tus of biomedical research personnel. This represents,to some extent, informed decisionmaking in the im-plementation of research training programs. The cur-rent number of trainees is quite close to the numberrecommended by the IOM committee.

Virtually all biomedical investigators have com-pleted training t.; the doctoral level. Nearly 60 per-cent of them go on to postdoctoral appointments, com-pared to 28 percent of the Ph.D. cohort in science andengineering in the aggregate. Allocation of Federal sup-port to either predoctoral training or postdoctoraltraining is a method of fine-tuning that can potentiallybe used to control the supply of biomedical investiga-tors. This mechanism is exploited only in the biomedi-cal sciences.

There appears to be no short-term shortage of basicbiomedical scientists, although there may be some tem-porary shortages of special personnel in biotechnol-ogy. As inehr.ry innovates in biotechnology, there willbe temporary shortages of highly specialized biomedi-cal research personnel, such as bioprocess engineers,molecular geneticists, and genetic engineers. In someareas of biomedical science, there may be a surplusof trained Ph.D. scienuts. While some Ph.D. biomedi-cal scientists may currently be having problems in ob-taining academic appointments, there is evidence thatthe rate of growth in industrial employment and self-employment may partially compensate for the surplus.Academic employment has bee.. increasing by morethan 4 percent per year, industrial employment is in-creasing by 8 percent, and self-employment is grow-ing the fastest, at nearly 25 percent. While the post-doctoral appointment has increasingly compensatedfur the lack of academic appointments, it is a short-term solution to the academic surplus that requiresmore extensive examination. Postdoctoral appointees,

do, however, conduct significant and important re-search in U.S. research institutions.

There is, however, evidence that the Nation is fac-ing a potential shortage of clinical investigators, asfewer physicians pursue research careers. As fewerphysicians enter into research careers, the potential fora widening gap between the direction of basic biomedi-cal research and the treatment of recognized diseasestates becomes greater. NIH has implemented fellow-ship programs to encourage medical students and newphysicians to pursue research careers.

Table A-2.Federal Manpower Legislation Pertainingto Biomedical and Medical Personnel

1930 Ransdell Act: created the National Institutes of Health.1937 Amendments to the Ransdell Act: established the Na-

tional Cancer Institute which established the first Fed-eral biomedical research program.

1944 Amendments to the Public Health Service Act: grantsauthority to NIH for extramural research programs.

1963 Health Professionals Education Assistance Act: estab-lished matching grants to assist in the construction ofteaching facilities for schools of medicine and healthprofessions, in response to the aggregate physiciansupply concem. Initiated a student loan program.

1965 Medicare/Medicald: created new demand for medicalcare, and further pushed the drive for increased biomed-ical research.

1985 Amendments to Health Professior.als Education As-sistance Act: extended the construction and studentloan provisions of original act. Established grants tohealth professions schools that increase enrollments(capitation).

1968 Allied Health Professionals Personnel Training Act:authorized establishment of a health professions stu-dent loan revolving fund.

1968 Health Manpower Act: extended the construction, basicand special improvement grants program, studentloans, and scholarship programs. Institutional grantsdetermined on a capitation basis.

1971 Comprehensive Health Manpower Training Acl: estab-lished four areas of authorizationconstruction grantsfor teaching and research facilities, capitation grantswith an emphasis on certain curricular components,student assistance, and computer technology healthcare demonstration programs.

1972 Emergency Health Personnel Act Amendments: Pub-lic Health Service and National Health Service CorpsScholarships were established with awards contingenton agreement to serve in a shortage area.

1974 National Research Service Award Act: abolished andconsolidated previous research Veining authorities un-der the Public Health Service Act.

1976 Amendments to the Comprehensive Health Manpow-er Training Act: capitation contingent on federally de-termined proportion of residencies in primary care.

1981 Omnibus Heconciliation Act: budget cuts in educationin the health professions, elimination of capitationgrants.

SOURCE Off Ice of Technology Assessment

Appendix B

Responses to an OTA Survey onInformation Needs Related to Minorities

in Science and Eingineering

The literature on the factors that affect minority par-ticipation in science and engineering is not nearly scwell developed as that for women. To supplement thefew comprehensive studies on this subject to date,OTA sent a questionnaire on "information needs re-lated to minorities in science and engineering" to 40recognized experts in the field. Respondents were askedto present their views as to the causes of and remediesfor problems in minority participation is science andengineering; the effectiveness of existing interventionprograms to promote such participation; and the needfor further research, additional information, and pol-icy actions. Responses received from 18 individuals aresummarized below, by question.

1. What do you believe to be the principal factors thatinfluence minorities' decisions to participate and con-tinue in science and engineering careers? Are there spe-cial factors that discourage minorities from partici-pating?

The negative factors most commonly mentioned byrespondents were:

lack of academic preparedness in elementary andsecondary school (literacy and necessary scienceand mathematics courses);lack of role models, mentors and teacher en-couragement;

a lack of parental support and encouragement;lack of peer support;inadequate career and academic counseling;lack of confidence and perception of self;financial strains; andlack of awareness of career opportunities.

Other negative factors included:societal emphasis on sports, rock stars, and"quickie" models of success rather than a slow andsequential model;loss of interest or motivation;poor study habits;lower educational and financial background(socioeconomic class standing);dry unimaginative teaching;lack of institutional commitment to minoritystudents;

7 47

declining number of qualified teachers and lackof in-service training opportunities for teachers;increasing remedial classes;identity problems (one respondent believes thisis especially so for American Indians);lack of summer jobs in science/engineering;lack of cultural and society support for science;

' type and environment of undergraduate insti-tution;lack of effective instructional programs to pro-mote cultural awareness and development ofbilingual skills (especially Chicano, Puerto Rican,and American Indian); andlack of transitional instructional programs for stu-dents with limited English language skills.

Positive factors for all minorities are:competence in the English language;early enrollment in math and science courses;continuation of math and sciegre sequence in sec-ondary school;basic interest in science and math;intervention programs;encouragement and support from mentors, fam-ily, and teachers;role models;positive input from peer group with high expec-tations;availability of financial resources;self-discipline;

e good study habits;continued success;challenge,good environment;intellectual gratification;opportunity to obtain research experience;second-generation college student (for blacks);awareness of careers and opportunities in scienceand engineering;starting salaries and possibilities for promotion;andone respondent said the perception of a scienceor math career as an a mnue for "escape" froman undesirable environment.

141

142

2. Are there major gaps in information and under-standing relative to the factors which influence minor-ity participation in science and engineering? If so,please identify.

A majority of the respondents felt that more infor-mation is needed on the experiences of minority stu-dents who are successfully participating in science andengineering. No major study exists on how graduatessucceed in completing science and engineering pro-grams. Why not study those who made it, let themexplain why, how, and where they came from beforethey completed a science or engineering degree? Studyevery member of several black, Hispanic, and Amer-ican Indian science and engineering societies to deter-mine what they have in common, and what theirdifferences are.

We do not understand the relative influence of thevarious factors that affect the participation of differ-ent minority groups in science and engineering. Littleattention has been given to differences in cognitivestyles that might exist for different minority groups.Many pieces of data, otherwise " available," are notbroken down by race and by sex, so that it is difficultto make full use of collected information.

Successful programs have not been evaluated to de-termine how they work, why they work, and whatthey have produced. Further research is required todevelop new strategies and programs that will en-courage minority students to pursue careers in scienceand engineering.

3. What new research initiatives would you recom-mend to help provide the needed information or un-derstanding relative to the factors influencing minor-ity participation in science and engineering?

One of the most frequently cited recommendationswas the need to improve the quality of mathematicsand science programs at the pre-college level, includ-ing improving basic skills. Improving the quality ofteaching was also highly recommended, and a needwas expressed for approaches that would motivate, up-date, and retrain inner-city school teachers. Guidancecounselors should be retrained and attuned to currentopportunities in science and math.

Research is needed on the outcomes of programsspecifically designed to recruit and maintain minor-ities in science and engineering. We know relativelylittle about which programs have succeeded and why.Previously supported programs that have had a strongpositive educational impact should be examined forposAble reinstitution.

Financial aid is still crucial to enable minorities toattend and remain in college.

There is a need to track students through the educa-tional pipeline; students who show an aptitude for

science and mathematics as early as junior high school.Participants in pre-college special programs should betracked for 10 years after high school graduation todetermine the relationship between program partici-pation and academic performance, retention in collegeand career choice. The relationship between careeraspirations of minorities between grades K and 12 andtheir decision to participate in science and mathematicseducation should be examined.

Research on the effect of role models and early ex-posure to science on career choice is needed.

Current information should be made available, bro-ken down by race and sex. More efforts are neededto sort out the relative weight of the differen: barriersto minorities' participation in science and engineering.

The whole web of cultural and social values shouldbe examined in order to understand why Asians andAsian Americans as a group have successfully enteredquantitatively based fields in disproportionately largenumbers relative to their representation in the popu-lation, in spite of numerous barriers. We need to iden-tify factors at work in the case of individual non-Asianminorities, who successfully enter science and engi-neering fields.

One respondent suggested the following additionalstudies: The family and educational backgrounds, andcareer pattern and experiences of minority scientistsand engineers. Quantitative studies of high school stu-dents experiences in mathematics and science courses.The attitudes of teachers and parents regarding appro-priate candidates for science/engineering. A study ofguidance counselors' familiarity with opportunities forblacks in science and engineering. More systematicevaluation of intervention programs. A study of theimpact of the availability of financial aid on careerdecisionmaking. A study of high school students' ca-reer awareness with respect to opportunities in science/engineering.

As a final note, three respondents believed no newresearch was necessary. What is needed is action.

4. In the recent past, a variety of programs and pol-icies have been initiated, on the Federal, State, local,and institutional level, to promote minority partici-pation in science and engineering. How effective havethe programs and policies with which you are mostfamiliar been in achieving their goals? Can you citespecific examples of success or failure?

Three programs were cited most often as successes:The Mathemat: t..s, Engineering, Science Achievement(MESA) ....rogram of the Lawrence Hall of Sciences inBerk 'ley, CA; the Philadelphia Regional Introductionfrir Minorities in Engineering (PRIME) program; andthe Resource Centers for Science and Engineering. TheAtlanta University Resource Center's Summer ScienceEnrichment Program was also mentioned. Only a few

I(.3

143

of the respondents hi.d actual knowledge of the pro-grams; most relied on hearsay.

In terms of actual experiences one respondent stated,I created the first successful program to bring blacks

into engineering in circa 1965-1968 at Oakland Univer-sity, Rochester, Michigan. In 1973 I began a programto bring women into Engineering at the University ofVirginia. We went from zero to the highest percentageof women in the Nation in five years.Another respondent reported first-hand knowlege

of the following minority programs sponsored by Fed-eral agencies: Minority Research Initiation (MRI), Re-search Initiation in Minority Institutions (RIMI), Mi-nority Biomedical Research Support (MBRS), MinorityAccess to Research Career (MARC), Research Centerfor Science and Engineering (RCSE), and the Minor-ity Institutions Science Improvement Program (MISIP).Of the preceding programs, only the Resource Centerswere conceived and planned to be comprehensive inapproach. Through the respondent's experience withthe Puerto Rico Resource Center, the respondent con-cluded that the only effective way to upgrade scienceeducation for minorities was to follow a holistic ap-proach in which resources are optimized to addressdifferent stages of the educational process. Most otherprograms tend to be targeted, rather narrow, andlimited in scope.

The same respondent stated that the MBRS andMARC programs have a strong overlap in goals, butnonetheless have been extremely successful in en-couraging and orienting students toward biomedicalresearch careers. The National Science Foundation(NSF) should be encouraged to consider the possibil-ity of creating equivalent programs in the physicalsciences, mathematics, and engineering. The MRI atNSF has been particularly successful in starting youngminority scientists in their research careers. Competi-tion for research funds is so keen that without the MRIprograms, many minority scientists might not be ableto establish themselves as regular faculty members inresearch institutions. Lastly, the RIMI program hasbeen crucial in providing infrastructure tur researchin minority institutions although the funds availableare not sufficient to remedy the seriousness of theproblem.

Other exemplary programs cited were: Project Act101 (Drexel University), Project Yes (University of theDistrict of Columbia), Southeastern Consortium forMinorities in Engineering (SECME), A Better Chance,Inc. of Boston, National Action Council for Minor-ities in Engineering (NACME), Phillips Academy, andthe Andover-Dartmouth Urban Institute intensive 4-week summer math matic program for second? -vschool teachers.

The most successful programs have been those withclearly defined goals; stable financial support and sta-ble staffing; a range of support services; and strongties to parents, professional societies, the local schools,and other community groups.

One respondent stated it this way: Most of the long-es,ablished programs are quite successful, but mostsuffer from inadequate long-term funding, inadequatelong-term assessment, little funding for transfer of suc-cessful models, and the need to prove themselves overand over again. The Resource Centers were a success-ful program at NSF, but they are not being supportedin 1985 because of the lack of long-term Federal com-mitment. The decision to discontinue the program wasmade independent of the fact that it works.

.mother expressed disapproval:Most of the Federal, State. or local programs have

not been the most effective use of resources. Programshave disappeared at the end of funding. Resourcesduplicated existing programs. California's investmentin MESA and Washington's in the minority engineer-ing effort, are examples of how to build results-oriented,long-term programs that have multi-sector support andwork with educational institutions for the benefit ofparents. Many of the old NSF programs were effectivein training teachers, but they are ended.

5. What have we learned from these programs and poi-ides; what specific factors appear to account for differ-ences in relative effectiveness?

Evidence obtained to date from the many interven-tion programs indicates that women and minority stu-dents will do as well as white males when providedwith quality education and support for their goals. Theearly identification of minority students with an apti-tude for mathematics and science, followed by a long-term enrichment and motivation program, has beenshown to he an effective approach.

Specific factors cited by one respondent as contrib-uting to the success of such programs were:

an effective liaison between primary and second-ary educational systems and higher education;parental awareness and participation;length of enrollment in the pre-college program;adequate re' urces; anda career development component.

Activities with a large motivational content, wherestudents are exposed to role models and to nontradi-tional ways of teaching science, were said to have aprofound and permanent effect in encouraging stu-dents to continue careers in science. Special programsto train kindergarten to sixth grade teachers in the pres-entation and teaching of mathematical and scientificconcepts, and the development of quantitatiw skillsand scientific methods were cited as crucial if minor-

7 4 9

144

ity students are to be oriented toward science andmathematics programs in large numbers. Respondentsstressed the importance of continuity in minority pro-grams. They must be comprehensive and maintainedover a sustained period of time.

One respondent stated that the principal lessonlearned from the Resource Centers was that a systems'approach is crucial to establish a comprehensive andcoherent plan to develop minority institutions into re-search centers and increase the number of students whogo on to become research scientists and engineers.

Other factors cited were:hands on experience;strong directors and able, interested staff;opportunities for in-school and out-of-schoollearning experiences;community support;development of a peer support sytems;evaluation;long-term follow up; andcareful data collection, and mainstreaming.

6. What additional information is needed to better un-derstand the causes of success t -.1 failure in programsand policies that promote minority participation inscience and engineering?

Most of the respondents felt that program evalua-tion is the most needed supplement to existing in-formation. This includes eva'-tation of interventionprograms and policies for promoting minority partici-pation in science and engineering; all pre-college engi-neering, mathematics, and science programs; and cur-ricula of primary and secondary systems attended byminorities compared to 'hose attended by majority stu-dents.

Further study and support of the following was alsosuggested: The current and future role of predomi-nantly minority institutions in the preparation of stu-dents for quantitative fields, the presence and role ofminorities in arenas where major science policy issuesand decisions are being made, successful efforts at theundergraduate level to recruit and retain minority stu-dents in quantitative fide:, specific ways to effectivelycommunicate with the minority community about theimportance of science and its applications in their lives,and the availability of computers to minority studentsand the ways in which they are being instructed to usethem.

Three of the respondents felt that no more informa-tion is needed, only action; the implementation of rec-ognized successful programs in education.

7. What actions, if any, would you recommend Con-gress take to promote minority participation in scienceand engineering, or to develop a better understand-ing of the factors that influence minority participation?

Most of tisi.- actions recommended for congressionalconsideration fell into three main categories: The estab-lishment and support of special programs for minor-ities in science and mathematics, strengthenirg thequality of mathematics and science educatio., . rt-eral, and financial support for minority studen, .

Some of the types of programs cited as deservingof support were: programs to promote curriculum de-velopment in mathematics and science for minoritystudents at the primary and secondary school level;programs that provide elementary and high school stu-dents with hands on exposure to science and engineer-ing; programs that encourage minority scientists andupper level minority science students to participate inminority science education; community-based pro-grams that utilize outside expertise, but maintain con-trol within local schools; programs aimed at increas-ing the number of students completing Algebra 1 andBiology 1 by the 9th grade, and the number of stu-dents taking and completing 3 or 4 years of mathe-matics, science, and English by the time they receivetheir high sci -1o1 diploma; and the development ofmagnet schools for the education of students with aspecial interest in science. It was recommended thatthe Government provide summer enrichment pro-grams for minority students with talent and interestin science we mathematics, and reestablish the Re-source Centers program.

Associated with program support was the need toimprove instruction in science and engineering in gen-eral. Pre-college science and mathematics instructionin urban schools needs to be improved with an em-phasis on involving both sexes. There should be in-creased funding for research on effective teaching andlearning of mathematical and scientific concepts andskills, with a reasonable proportion set aside to studypopulations at risk for low achievement on the fun-damental cognitive and conceptual level. We shouldcontinue informing the very young of opportunitiesin science and engineering, and educating the educa-tors. Finally mathematics and science education in gen-eral should be strengthened through special fundingfor equ.,....aent and teacher training.

Financial support for minority students was sug-gested via various means: Establishment of "congres-sional scholars in science and technology" (a com-prehensive 3- or 4-year program of undergraduate

1 v0

145

support, research, and work experience to ensure thattalented minority students will not be lost to scienceand engineering because they are unable to pay thecost of first-rate college training in these fields). Two-year scholarships could be awarded to a group of-r- nority high school juniors, and to a comparablenumber of minority college juniors, who have dem-onstrated significant interest in and potential for suc-cessfully pursuing science or mathematics .tudy at thenext educational level. (Consideration might b' --;vento targeting some of the proposed awards ' entsinterested in pursuing science/mathematics t Achingcareers.) Financial support to students could be in-creased, especially support not involving loans. No-interest or low-interest college loans could be providedto students with strong mathematics and science back-grounds who agree to teach mathematics and sciencecourses in economically disadvantaged communities.A federally guaranteed, low interest loan programcould be established and specifically designated for minor:ties who indicate an interest in becoming scientistsor engineers.

0

Some respondents recommended an examination ofthe Department of Education and NSF programs forminorities to determine their effectiveness, and theneed for additional financial support. Continuity ofsuccessful minority programs shouiu be the guidingfactor in the determination of funding by Congress forminority programs, in their view Such programsshould be "mainstreamed" into the '.2gular educationalprocess, where possible.

A number of respondents called for a centralizedorganization to deal with tie pn ?terns of minoritiesin science and mathematics. A nati9nal conference orhearing could be convened on min ,rity participationin science and engineering. A commission could beestablished, composed of scientists and educators fromthe underrepresented groups as well as representativesfrom employers affected by the underrepresentation(e.g., universities, industry, national labs). A nationaloffice is needed, acr -no to some respondents, to fa-cilitate 7 id cond, ..,c activities designed to increase mi-nority , , ticipation in science and mathematics.

Office of Technology Assessment

The Office of Technology Assessment (OTA) w7 3 created in 1972 as ananalytical arm of Congress. OTA's basic function is to help legislative policy-makers anticipate and plan for the consequences of technological changesand to examine the many ways, expected and unexpected, in which tech-nology affects people's lives The assessment of technology calls for explora-tion of the physical, biological, economic, social, and political impacts thatcan result from applications of scientific knowledge. OTA provides Con-gress with independent and timely information about the potential effectsboth beneficial and harmfulof technological applications.

Requests for studies are made by chairmen of standing corn: uttees ofthe House of Representatives or Senate; by the Technology Assessmen.Board, the governing body of OTA; or by the Director of OTA in consul-tation with the Board

The Technology Assessment Board e, composed of six members of theHouse, six members of the Senate, and the OTA Director, who is a non-voting member

OTA has studies under way in nine program , reas energy and materi-als, industry, technology, and employment, international security and com-merce, biological applications; food and renev:able resources, health;communication and intormation technologies, oceans and environment; andscience, education, and transportation

I

OTATM-SET-35 DECEMBER 1985

I -... t 0