Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 2 THE CURRENT MARKET FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS OVERVIEW Labor markets for biomedical and behavioral scientists moved toward a balance between supply and demand in the 1980s after a period of excess supply in the 1970s. Biomedical scientists are in strong demand due to increased employment in industry; increased predoctoral enrollments have not yet produced an adequate supply of new biomedical Ph.D.s. The behavioral sciences have worked off the excess supply of the 1970s through continued employment growth and a decline in the number of new Ph.D.s. There are no reliable data on the supply and demand for physician/scientists. Despite recent progress, minorities remain underrepresented among Ph.D. recipients and in the total work force of the biomedical and behavioral sciences. Female participation has increased more rapidly than minority participation, particularly by black males. However, many female Ph.D.s are not employed full-time in their scientific fields, with adverse consequences for personnel supply. Foreign students received only 18 percent of biomedical Ph.D.s and 7 percent of behavioral Ph.D.s in 1988. However, most foreign students do not stay to work in the United States after graduation. Thus, foreign students contribute relatively little to the growth of biomedical and behavioral sciences work force. THE CHANGING LABOR MARKET FOR BIOMEDICAL AND BEHAVIORAL SCIENTISTS, 1973-1987 The size and composition of the scientist work force . · · . . . ~ are determined by three flows: new entrants, usually new degree recipients; attrition, in the form of deaths and retirements; and net gains or losses due to occupational mobility. New entrants serve the dual purpose of offsetting attrition and permitting growth. The adequacy of supply of new entrants depends on whether their numbers exceed, equal, or fall short of the replenishment needs and growth needs of the work force. As a result, the accurate measurement of these three processes is basic to any study such as this. Refreshment Rates Figure 2-1 shows the attrition and refreshment rates for biomedical scientists--that is, the number of new Ph.D.s as a percent of those employed in that field. Attrition rates are the percentage losses due to death, retirement, and net mobility.2 Refreshment minus attrition is the percentage available for growth in employment in the field. In 1974, for iObviously, not all new Ph.D.s in biomedical science go into the field; also, the field draws Ph.D.s from other areas (e.g., physical sciences, other life sciences) and from foreign scientists. However, the primary supply source of new Ph.D.s is U.S. graduates in the field, and thus the refreshment rate gives one a sense of the historical relationship between this supply source and demand. 2Net mobility is the difference between scientists leaving the biomedical field for other pursuits and scientists entering biomedical science from other employment. 33
example, new Ph.D.s comprised 8.6 percent of total employment in the biomedical sciences, while losses from attrition were 2.6 percent of employment; this left a net of 6.0 percent available for growth in employment of people trained in the field. By 1987, however, refreshment was S.4 percent, attrition 2.8 percent, and the potential ' for growth only 2.6 percent. Actual employment growth, 4.1 percent that year, was achieved by drawing new entrants from other fields. During the 1980s, the biomedical field averaged 4,500 job openings annually (1,080 ' scientists lost from deaths and ' retirements, 620 scientists lost from net mobility, and growth requirements of 2,800~. Average annual biomedical Ph.D. production during the period was only 3,840. This contrasts sharply with the 1970s, when average annual job openings from all sources were 3,660 and average annual Ph.D. production was 3,520. Clearly, the supply of new biomedical Ph.D.s has begun to fall short of the number of job openings in the late 1980s after an extended period of anoroximate balance. In the behavioral sciences. On the other hand, new Ph.D.s have exceeded new job openings ~` throughout the 1970s and 1980s. a s 4 3. 2 1974 1975 1976 1977 1978 1979 1980 19811982 1983 1984 1985 1986 1987 Year SOURCE: Appendix Table A-~. Figure 2-1. Refreshment and attrition in biomedical science, 1973-1987. Figure 2-2 shows that the ~ 12 refreshment rate for behavioral ;-, scientists (excluding clinical ° psychologists) has also been ~ 10- decreasing. In 1973 attrition ~ amounted to 4.4 percent of those O employed and refreshment was _ 13.S percent, leaving 9.4 percent for growth; actual growth was lo, 8.6 percent. In 1987, attrition was 5.4 percent, and refreshment was 7.0 percent, leaving 1.6 percent for growth; and indeed actual growth was only 1.2 percent. Attrition Figure 2-3 presents estimates of the rate of exit from the science work force due Figure 2-2. Refreshment and attrition in nonclinical to death and retirement. These psychology and other behavioral sciences, 1973-1987. _~ Attrition Rate `` ~ Refresh. rate 8- 6 ~ _\ of__ . 1974 1975 1976 1977 1978 1979 19801981 1982 1983 1984 l9B5 1986 1987 Year SOURCE: Appendix Tables A-9 and A-10. 34
rates increase sharply for scientists over 55 years of age.3 Net Mobility Scientists who leave biomedical science for other employment must usually be replaced.4 Figure 2-4 shows estimates of these net rates of mobility (outmobility less inmobility from other fields) based upon historical data.5 These exit rates are a function of scientists' "career age" (years since Ph.D. degree); Figure 2-4 shows that the rate increases as a scientist matures. Obviously, scientists from a wide variety of degree fields (e.g., physical science, behavioral science) work as biomedical scientists. The net mobility rates used here are based on field of employment, regardless of the scientist's degree field. Once scientists gain employment either in a biomedical field or in a nonbiomedical field, they are identified by that employment field rather than their field of degree. RECENT TRENDS IN THE LABOR MARKET FOR BIOMEDICAL SCIENTISTS Figure 2-5 displays patterns of employment for biomedical scientists in 1987 by employment sector and by primary work activity--A&D or ~ 0.20- c L4 - 0~10 - ~5 _! f 0.00- _~ , . .. <30 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 Biological Age SOURCE: Appendix Table A-16. Figure 2-3. Death and retirement rates for scientific employment. - 0.06 - 0.05 - a,- 0.04- a_ i_ 0 - 0.03- `,, - 0.02- z ~ . . Total Scientist · R&D Scientist I ~ nor <pi 6-10 11-15 16-20 21-25 26-30 31-35 36-40 Career Age (Years since Ph.D.) SOURCE: Calculated from Appendix Table A-17. Figure 2-4. Net mobility rates for total and R&D biomedical scientists by career age. 3For a more detailed discussion of the model used to estimate scientist attrition, see Joe G. Baker, "Biomedical/Behavioral Cohort Model: A Technical Paper," contained in volume III of this report. 4Biomedical scientists are lost from the work force when they move into university administration, industrial management, or other non-science occupations. This loss is partially offset by individuals who move into biomedical science from non-science. The net mobility estimates used here are the difference between these two labor flows. 5The rates in Figure 2-4 are for biomedical scientists only; separate rates were computed for behavioral scientists. 35
40000 30000- it - 20000 10000 - Academic Postdoc SOURCE: Appendix Table A-2. ~ . 11 .... . . . 1~ ! 1 ~ o 11i 11 ~ 110~: Non-it&D Empl. ~ 1 RED Empl. Industry Government All Other Figure 2-5. 1987 biomedical science Ph.D. employment, by R&D activity and employment sector. the management of R&D, and non-R&D.6 An estimated 76,300 Ph.D. scientists identified their work as biomedical science, and another 8,200 were undertaking postdoctoral study in the biomedical sciences. Academic employment (43,000 scientists) and industrial employment (16,000 scientists) were the largest sectors. Overall, 60.5 percent of all biomedical scientists indicated that their primary work activity was R&D or the management of R&D. Growth in Employment Figure 2-6 shows the growth in employment of biomedical scientists between 1973 and 1987. Total employment plus postdoctorates nearly doubled, from 43,000 in 1973 to 84,500 in 1987, an annual rate of growth of 4.9 percent. Industrial employment increased at over twice the rate of academic employment. The proportion of biomedical scientists who indicated that their primary work activity was R&D or the management of R&D also 6In the Survey of Doctorate Recipients (SDR), respondents are asked to identify their primary work activity. In this report, those respondents who identified their primary work activity as R&D or the management of R&D are classified as R&D scientists; all other scientists are classified as non-A&D scientists. Obviously, some (or most) of these non- R&D scientists devote some portion of their work activity to R&D; however, it is not treated as their primary activity. The R&D/non-R&D dichotomy is used to indicate differences in research intensity among employment sectors and through time. 36
10 8 - lo ~ 0 - I: ¢ ED - ¢ 4 2- o ~ R&D Empl. · Total Scientist il Academic Postdoc lddus~y Govern. All Other Total SOURCE, Appendix Table A-2. Figure 2-6. Average annual growth rates for R&D and total biomedical Ph.D. employment, 1973-1987. increased during this period, from 53.6 percent in 1973 to 60.5 percent in 1987.7 This was the result of two basic trends: a growth in the proportion of total employment made up of private industry, in which more than three-fourths of biomedical scientists are engaged in R&D; and an increase in the percentage of academic scientists who indicate that their primary work activity is R&D. The annual average increase in the number of biomedical scientists engaged in R&D has been 5.S percent for the 1973-1987 period, slightly higher than overall growth in the field. Postdoctoral Appointments Figure 2-7 shows that the number of postdoctoral training positions in the biomedical sciences grew rapidly through the 1970s, but has since plateaued~ near the 1981 7These estimates are from the NAS/NRC, Survey of Doctorate Recipients (SDR). 37
level.8 These trends are consistent with the evidence that new job openings exceeded new Ph.D.s in the late 1980s. Trends in Job Openings and Ph.D. Production Using the death and retirement rates from Figure 2- 3, one can estimate attrition from the worl`' force of - biomedical'scientists based on age distribution. For the period 1973- 1977, deaths and retirements from the stock of biomedical scientists averaged approximately 730 per year. In the later period'l983-1987', the number of annual deaths and retirements averaged 1,200, a 64- percent increase over the earlier period. 9000 ~ a . - o 6000- 8000 7000 o L,, 5000 in 4000 - .. 1973 1975 1977 1979 ~ 1981 1983 1985 1987 SOURCE: Appendix Table A-2. Figure 2-7. Postdoctoral appointments of biomedical scientists, 1973- 1987. For the period 1973-1977, there was a net loss of approximately 450 biomedical scientists per year to other employment. These losses were replaced by new hires. Because of employment growth and changes in the career age distribution, this net migration grew to approximately 670 per year in the period 1983-1987. By the late 1980s, therefore, the need to replace attrition with new Ph.D.s had grown substantially. The aggregate annual attrition of approximately 1,200 consumed one-third of the yearly biomedical Ph.D. production of 3,500 Ph.D.s during the 1973- 1979 period. For the period 1979- 1987, annual attrition grew to 1,700 and consumed almost half of the annual biomedical Ph.D. output of 3,850.9 Indeed, at current rates of employment growth and Ph.D. output, attrition replacement will equal new biomedical Ph.D. output by the end of the century. Figure 2-S compares the number of job openings by source with the annual production of new biomedical Ph.D.s for the period 1973-1987. It shows a growing gap between supply of and demand for new Ph.D.s. Biomedical science Ph.D. awards totaled 3,520 in 1973; this grew to 3,960 in 1982 and has been relatively-flat since then. As a result, there was approximately one job opening per new Ph.D. for the period 1973-1979, but for the period 1979-1987 there were on average 1.17 job openings for each new Ph.D. There are two sources of data for postdoctoral appointments in the biomedical sciences: the SDR and the National Institutes of Health/National Science Foundation's Survey of Graduate Science and Engineering Students and Postdoctorates (GSESP). 'The SDR data are used here. 9Because death and retirement are a function of biological age and net mobility is a function of career age, the overall attrition rate of the work force is related to its career and biological age structure. In the 1980s, death and retirement were approximately 1.9 percent per year and net mobility approximately 1 percent per year. See Joe G. Baker, op cit. 38
Biomedical Scientists ~2 o 2000 vat or ' ~ 6000- ~ 5000 - 4000 - 3000 - 1000 - O- ~ ~ 1973- OP~g Ph.D.s , ~ - d ~ 61] 1 IN SOURCE: Appendix Table A-~. Op~gs 1979 . 1979^ 2l PhDs 1987 [23 Ph.D.s Grown eel Net Mobility ~ DeathlReiired Figure 2-8. Average annual job openings and new Ph.D.s in the biomedical sciences, 1973-1979 and 1979-1987. People and institutions seem to have responded to this imbalance between job openings and new Ph.D.s in the biomedical sciences.~° Real wages increased for biomedical scientists in the late 1980s, as they did for all Ph.D.s. In response, graduate enrollments in the biomedical sciences (both doctoral and master's levels), which had declined to a low of 41,191 in 1983, have since grown steadily to 44,495 in 1987. First-year graduate enrollments in doctorate-granting institutions have risen from only 8,043 in 1983 to 8,609 in 1987. However, the enrollment response requires several years before numbers of new Ph.D.s are affected. RECENT TRENDS IN THE LABOR MARKET FOR BEHAVIORAL SCIENCES Behavioral scientists are composed of three basic groups according to the NRC nomenclature: clinical psychologists, nonclinical psychologists, and other behavioral scientists (anthropologists, sociologists, audiologists, and speech pathologists). Clinical Psychology Psychologists, who form the majority of behavioral scientists, have available an alternative to the standard research career pattern. This consists of independent practice in the broad areas of clinical and counseling psychology. Concerned with patient care Lessee Joe G. Baker, "The Ph.D. Supply Crisis: A Look at the Biomedical Sciences," paper given at the Western Economics Association Meeting, June 21, 1989, Lake Tahoe, Nevada. Lithe supply consequences of increasing enrollment were partially offset by increasing time to complete doctoral studies. For a more complete discussion of increasing time to the doctorate, see H. Tuckman, et al., On Time to the Doctorate, (Washington, D.C.: National Academy Press, forthcoming. See also Joe G. Baker, "The Ph.D. Supply Crisis." 39
rather than research, this-career path resembles that of the practicing M.D. Doctorates in psychology are usually obtained from university departments, although professional schools of psychology are being accredited to provide doctorates in clinical and counseling fields that are oriented toward service and/or administration. Only about 1,200 out of a population of 33,388 clinical psychologists (less than 4 percent) are engaged primarily' in R&D or the management of R&D. Given the applied nature of clinical' psychology, this report focuses on nonclinical psychologists and other behavioral scientists. Nonclinical psychologists usually will be refer-red to simply as psychologists hereafter. Nonclinical Psychology The market for psychologists is dominated by the academic sector, which employed 63.7 percent of the 1987 total of 20,510 psychologists (Figure 2-9~. Industry employed another 16.5 percent, and the remaining 20 percent of employment was scattered across other sectors. In 1987, 29.9 percent of all psychologists indicated that their primary work activity was R&D or the management of R&D. Industrial employment grew at over twice the rate of the academic sector between 1973 and 1987 (Figure 2- 10~. Nonclinical psychology degree holders have also been attracted to employment in clinical psychology.~3 Employment of psychologists increased at an annual rate of 3.2 percent from 1973- 1987. Postdoctoral appointments in psychology are few compared to biomedical fields, increasing from 259 in 1973 to 666 in 1987. For the period 1973- 1979, an average of 230 psychologists retired or died annually and approximately 430 were lost each year to net mobility. The sum of these two losses--660-- represented about 41 percent of annual new nonclinical psychology Ph.D.s ~ 1,591). For 12000 ~ - ~ 9000. - i~ to - ~3 6000~ ~ Non-it&D Empl. ED R&O Emol. 0 1 - 1- -- -I- - ~ - ,!-,. an,,. , , I, ,, Academic Postdoc Industry Government All Other SOURCE: Appendix Table A-3. Figure 2-9. 1987 Ph.D. Employment of nonclinical psychologists by R&D activity and employment sector. Din 1987, an estimated 421 Ph.D.s were doing postdoctoral work in clinical psychology. The employment growth of clinical psychologists was substantial for the 1973-1987 period, averaging 7.9 percent annually; but most of this growth was concentrated in the practice sector (where fewer than 0.1 percentile clinical psychologists are active in R&D). These data are from the Survey of Doctorate Recipients. loin 1987, 16.1 percent of all employed clinical psychologists had earned their Ph.D.s in nonclinical psychology, and another 5 percent came from other fields. The reverse movement, from clinical to non-clinical psychology, was almost identical--16 percent, according to the Survey of Doctorate Recipients. 40
the period 1983-1987, attrition from death and retirement grew to 380 annually, and mobility to 540; this aggregate annual attrition of 920 represented 62 percent of average new Ph.D. production of 1,485.~4 As shown in Figure 2-11, there were about 0.7 job openings for each new nonclinical psychology Ph.D. in the period 1973 1979, rising to approximately 0.9 job openings per each new Ph.D. in the period 1979-1987. The labor market of the 1980s for nonclinical psychologists thus seems more in balance, both as a result of increasing openings (primarily from attrition) and a decline in the average number of nonclinical psychology Ph.D.s produced, from 1,591 per year (1973-1979) to 1,485 per year (1979-1987~. Other Behavioral Sciences Employment in these fields increased at an annual rate of 4.S percent for the 1973- 1987 period (Figure 2-12~. However, the growth rate for the 1983-1987 period slowed to 0.3 percent annually, and academic employment actually declined during the period. Academic employment of other behavioral scientists dominates their labor market: in 1987 almost 85 percent of academic employment was in colleges and universities, with the remaining 15 percent scattered across other employment categories. In 1987, total employment in other behavioral sciences was 12,735 (excluding 192 postdoctorates). 8- - - In, 6- - - 0 4 - - ~ 2 I: 6 o . ~ :! _ ~ ._ 1 ~ e] R&D Empl. 1151 Total Empl. l Academic Postdoc Industry Govern. All Other Total SOURCE: Appendix Table A-3. - Figure 2-10. Average annual Ph.D. employment growth for R&D and total nonclinical psychologists, 1973-1987. . 2000. In 1500- OD 0 1000- a: I: see - 0~ ~ o- Nonclinical Psychologists Ph.D.s Ph.D.s Openings Fin - ~ 1973- 1979 . 1979- 1987 SOURCE: Appendix Table A-9. Other behavioral ~ Figure 2-11. Average annual job openings and new scientists engaged primarily in Ph.D.s in nonclinical psychology, 1973-1979 and 1979- R&D averaged about 9 percent i4For a more detailed discussion of the model used to estimate scientist attrition, see Joe G. Baker, "Biomedical/Behavioral Cohort Model: A Technical Paper," in Volume III of this report. 41 E:3 Ph.D.s 1~3 Grown Ed Net Mobility Ed DeadllRetire.
of total employment for the entire 1973-1987 period. Approximately 15 percent of those academically employed indicated that their primary work activity was in R&D. Postdoctoral appointments of other behavioral scientists comprised only 1.4 percent of the total work force in 1987. In recent years, the majority of job openings for other behavioral scientists have resulted from attrition. Death and retirement (130 average annual openings) and negative net mobility (250 average annual openings) resulted in approximately 380 job openings annually during the 1973-1978 period; new growth required almost 780 scientists annually. In the 1983-1987 period, death and retirement (280 openings) and net mobility (370 openings) both exceeded annual growth positions (approximately 220 new Ph.D.s). 14 - 12 - 10 - 8 - O ~ 2.'..~.2.2...~..... .,.,,,,,,.,,, Academic Postdoc Industry SOURCE: Appendix Table A-4. 3 R&D Empl. · Total Scientist All Other Total Figure 2-12. Average annual growth in Ph.D. employment of other behavioral scientists by total and R&D, 1973-1987. Degree production in other behavioral sciences fell steadily, from 1,307 Ph.D.s in 1976 to 882 in 1987, mirroring the decline in annual openings. The number of job openings per new other behavioral sciences Ph.D. had not changed substantially during the 1973- 1987 period, ranging from 0.85 to 0.95 openings per each new Ph.D. Scientists who hold Ph.D.s in other behavioral science fields have also sought employment in other fields as a result of this soft market: in 1987 only about half of all scientists with Ph.D.s in other behavioral science fields were working as other behavioral scientists, compared with 71.9 percent in 1972. RECENT TRENDS IN THE LABOR MARKET FOR PHYSICIAN/SCIENTISTS There is no precise tally of physician/scientists currently active in biomedical research, but indirect indicators point to a predominance of activities other than research. An example is a 1983 survey of full-time faculty within the departments of medicine at U.S. medical schools approved by the Liaison Committee on Medical Education (LCME).~5 Faculty in departments of medicine are traditionally viewed as being more involved in research activities than faculty from other clinical departments. The data from this survey can be used as an approximation of the upper limits of research effort for all full-time faculty. The survey found that 50 percent of the full-time faculty physicians spent less than 25 percent of their time conducting research; only 20 percent spent more than half of their time in research endeavors. This suggests that the expanding pool of full-time clinical faculty in medical schools (Figure 2-13) is a response to increasing patient care activities by clinical departments and does not reflect an increasing supply of i5H. N. Beaty, et al., "Research Activities of Faculty in Academic Departments of Medicine," Annals of Internal Medicine, vol. 104, 1986, pp. 90-97. 42
6000 - ca 5000- 0 4000- ~ 3000- :, as 2000 - 100(J - SOURCE: Appendix Table A-ll. <q~C of ~_\ - ~ Trmnee/Fellow ~ Faculty (in Ins) 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 Figure 2-13. Trends in M.D. postprofessional training and medical school faculty 1961-1987; faculty in tens (e.g., 1976 = 2S,603 faculty members and 2,918 trainees/fellows). physician/scientists. Extrapolating from this survey, it can be estimated that, at most, only 20-30 percent of the full-time faculty in all clinical departments are actively engaged in biomedical research. Several earlier studies also document the declining numbers of physicians entering careers in scientific investigation. As a case in point, the number of physicians in research training programs sponsored by NIH decreased, from approximately 4,200 in 1969 to 1,910 in 1977 and has not increased significantly since then (see Figure 2-13~.~7 This decline is explained, in part, by changes in program policy--much of the support in the 1960s and early 1970s was for training in medical subspecialities and required no research Resee, for example, S. O. Thier, et al., "Proposals Addressing the Decline in the Training of Physician Investigators: Report of the Ad Hoc Committee of the AAMC,~ Clinical Research, 1980, pp. 85-93, ]. B. Wyngaarden, "The Clinical Investigator as an Endangered Species," New England Journal of Medicine, vol. 301, 1979, pp. 1,254-1,259; and G. F. DiBona, "Whence Cometh Tomorrow's Clinical Investigators?", Clinical Research, vol. 27 1979, pp. 253-256. JIG. ~ Carter, A. Robyn, and A. M. Singer, "The Supply of Physician Researchers and Support for Research Training: Part I of an Evaluation of the Hartford Fellowship Program," report N-2003-HF, (Santa Monica, CA: The Rand Corporation, 1983), pp. 46-47. 43
focus. Regardless, the number of physician trainees has not kept pace with growth in the physician population. Data for NIH grant awards provide another indicator of declining research activity by full-time faculty with the M.D. degree versus those with a Ph.D. degree. Of 11,683 grants in 1970, 36.7 percent (4,289) were awarded to ~D. principal investigators, 51.3 percent (5,993) to Ph.D.s, 5.9 percent (693) to M.D./Ph.D.s, and 6.1 percent (708) unknown. By 1987, when 24,384 grants were awarded, 26.2 percent (6,393) went to M.D.s, 63.9 percent (15,589) to Ph.D.s, 3.7 percent (902) to M.D./Ph.D.s, and 6.1 percent (1,498) unknown. Although the number of grants to M.D.s increased, their share of total growth fell. COMPOSITION OF THE LABOR FORCE Race and Sex Over the past decade, those concerned with the scientific work force have researched and written extensively about the underrepresentation of women and minorities in that sector.~9 This concern is motivated by reasons of- both equity and strategy. But while the composition of the scientific work force reflects increasing numbers of individuals other than the traditional pool of white males, the participation of women and minorities in science is still far lower than their participation in the overall labor force . For instance, nearly 45 percent of the 1987 U.S. labor force was made up of women but only about 22 and 34 percent of biomedical and behavioral scientists were female.26 The distribution of minorities in science differs even more sharply from that in the larger labor force: blacks and all other minorities except Asians are underrepresented by factors of 6 or 7. In addition, while the percentage of women in the scientific work force is growing at a relatively rapid pace, the growth in the number of racial and ethnic minority scientists is painfully slow. An important trend in the general labor force also holds for biomedical/behavioral doctorate recipients: a much more rapid growth in the rate of entry for women than for men. Although the number of awards to minority males in biomedical science grew only slightly between the time period 1978-1982 and the time period 1983-1987 (3.61 percent), growth in the number of awards to minority women (27.5 percent) resembled that of white women. The trends were similar in the behavioral sciences, where female participation is i8B. Healy, "Innovators for the 21st Century: Will We Face a Crisis in Biomedical- Research Brainpower?," New England Journal of Medicine, vol. 319, 198S, pp. 1,059-1,064. resee, for instance, The White House Task Force on Women, Minorities and the Handicapped in Science and Technology, Changing America: The New Face of Science and Engineering, Washington, D.C.: September 1988; National Science Foundation, Women and Minorities in Science and Engineering, Washington, D.C.: U.S. Government Printing Office (a biennial publication beginning in 1982~; National Research Council, Women: Their Underrepresentation and Career Differentials in Science and Engineering, Washington, D.C.: National Academy Press, 1987; National Research Council, Minorities: Their Underrepresentation and Career Differentials in Science and Engineering, Washington, D.C.: National Academy Press, 1987; and Government-University-Industry Research Roundtable, Nurturing Science and Engineering Talent, Washington, D.C.: National Academy Press, 1987. 20Data are from National Research Council, Survey of Doctorate Recipients, 1987; and Council of Economic Advisors, Economic Report of the President 1989, Washington, D.C.: U.S. Government Printing Office, 1989. 44
I: s i, A: ~ 3 ~ 4) cat : ° 1 - - ~ -~ ~ ~ ~ 3 c ° ~ 5 _ ~ I- ~ -7 9 - traditionally higher, by the mid-l9SOs, over half of all behavioral Ph.D.s were being awarded to women. Comparable data are not available for physician/scientists. To determine whether there is an association between race or sex and the careers of biomedical and behavioral Ph.D.s, the committee examined predoctoral support, postdoctoral plans, and a career outcome of those in the doctoral labor force in the biomedical and behavioral sciences.2i Figure 2-14 indicates that the distribution of NIH predoctoral support to graduate students is not independent of race and sex in some cases. Both this figure and the two that follow show percentage point deviations of the race/sex groups from their respective field percentages. For instance, 23.3 percent of all biomedical degree recipients in the period 1983-1987 reported NIH as the primary source of support in graduate school. The comparable figure for black women in that group was 13.6 percent or 9.7 percentage points below the field percentage; in other words, black women were 9.7 · Biomedical Sc~endsts Off '/ ;/J ,~70 Behavioral Scientists ~~v,=~.~ 121 Men ~ Women _ I _ ~ ~ I ~ ~ - - ~ -11 Asian Black Other White Asian Black Other White SOURCE: Appendix Table A-12. Figure 2-14. Under- and over-representation of race/sex groups in NIH predoctoral support for biomedical (field percent = 23.3) and behavioral (field percent =- 2.9) doctoral recipients, 1983-1987. Tithe first group contains so few minority members in any given year that the committee, using the National Research Council's Survey of Earned Doctorates (SED), aggregated cohorts of doctorate recipients over two five-year periods to investigate both differentiation and change. For the same reason, it would have been desirable to aggregate across cohorts of the Survey of Doctorate Recipients (SDR), used for compilation of the second group, but the longitudinal nature of the SDR would have involved multiple counting of individuals. Hence, the two single years, 1977 and 1987, were selected for analysis of these populations. 45
percent underrepresented. The reasons for this are not known, however, and the numbers are small. Figure 2-14 shows that minority groups are uniformly underrepresented among basic biomedical doctorate recipients reporting NIH as their primary source of predoctoral support. In the behavioral fields, the situation is reversed, but the deviations are trivial. The underrepresentation of minority graduate students with NIH support is surprising, given major efforts to target minorities within NIH/ADAMHA. In addition to the MARC program, noted in the Executive Summary, NIH provides summer research apprenticeships for minority high school students, research grant supplements for projects that employ minority students or faculty, and strong directives to grant applicants and internal staff concerning procedures for recruiting minorities.22 Despite these efforts, the growth of the minority doctoral population in these fields has been slow. Possible explanations for the inconsistency between the size of the NIH/ADAMHA minority programs and low levels of actual minority student support include the following: 0 o o The major mechanism for predoctoral support within NIH/ADAMHA is the T32 institutional training program. The selection of individual applicants in these programs is left up to the institutions. NIH/ADAMHA requires schools receiving T32 grants to have a minority recruitment program, it is not known how effective the individual institutions are in recruiting minorities nor how under;represented minorities may be in the institutions. The NIH/ADAMHA minority programs may simply be ineffective. For example, if minorities have higher attrition rates and longer time to degree than other students, fewer Ph.D.s are produced per given level of predoctoral support. Thus NIH/ADAMHA support for predoctoral minority students may be high, but relatively few of these students may receive Ph.D.s. The inconsistency between NIH/ADAMHA minority programs and actual levels of minority student support may be due to data bias. Students responding to the Survey of Earned Doctorates may not be sure what their source of support was during graduate school.23 The causes of the low levels of minority student support and Ph.D. awards are clearly a cause for concern. The committee recommends program evaluations and pipeline studies as a start to sort out the causes of these troubling statistics (see recommendations in chapter 5~. To learn about postdoctoral plans, the committee relied on the SED, which asks whether respondents have definite plans for future study or employment at the time of completing the survey form, usually when the dissertation is accepted, those who answer negatively are displaying ambiguity about their careers. Figure 2-15 displays the levels of such uncertainty and any possible differentials by field, race, and sex; every measured effect operates in ways that are highly significant, both statistically and substantively. For example, about 9 percent fewer behavioral than biomedical majors report definite plans. A 22For details, see NIH Guide for Grants and Contracts, Special Issue: Initiatives for Underrepresented Minorities in Biomedical Research, vol. IS, no. 14, April 21, 1989. 23There are indications that minorities in the life sciences are receiving federal support from all sources at rates comparable to other groups. Those indicating primary support from any federal source were-93.& percent overall, 20.0 percent for American Indians, 43.7 percent for Asians, 31.3 percent for blacks, 29.5 percent for Hispanics, and 26.9 percent for whites. See National Research Council, Summary Report 1988: Doctorate Recipients From United States Universities, Washington, D.C.: National Acaclemy Press,~ 1989, Table L. 46
10 - 6 _ ~ 2 -2 - ~ -6 O ~ -10 _ ~ ~ Q) - -14 -18 ^^ _._ ^^ - ~ Biomedical Scientists .~ _ Behavioral Scientists E3 Men e] Women Asian Black Other White SOURCE: Appendix Table A-13. Asian Black Other White Figure 2-15. Under- and over-representation of race/sex groups in reporting definite postdoctoral plans of biomedical (field percent = 88.5) and behavioral (field percent = 79.2) doctoral recipients, 1983-1987. larger fraction of whites report definite plans than do minorities; and in all categories other than black, fewer women than men report definite career plans. The committee believes that the data in Figures 2-14 and 2-15 are indicative of inadequate mentoring at the predoctoral level: role models for women and minorities are too few in number, and contacts with faculty may be too sparse to provide needed guidance in seeking NIH support in graduate school and in career planning. These factors are very likely compounded with others, such as inadequate precollege preparation. Whatever the specific mix of causes, the committee interprets the data as suggesting a clear need for future research and appropriate action. These patterns in degree awards are having predictable time-lagged effects on the composition of the biomedical and behavioral labor force. The data in Table 2-1 reflect an average annual growth rate of about 4.7 percent in the size of the total labor force between 1977 and 1987. During that time period, the percentage of women in both doctoral fields grew substantially, to more than 20 percent of the biomedical work force and more than one-third of the behavioral scientists. The average annual growth in the numbers of white females were 7.7 and 8.3 percent, respectively; these rates of increase were shared by nonwhite women and, to a lesser extent, by minority males. The only groups growing at less than these rates were white males, whose numbers increased by only 3.6 and 2.9 percent in the biomedical and behavioral fields, respectively. 47
TABLE 2-1: Doctoral Biomedical and Behavioral Labor Force, by Sex and Race, 1977 and 1987 . Bionedical Behavioral _ 1977 1987 1977 1987 MEN Asian 4.83 6.95 1.04 1.20 Black 1.00 1.01 0.78 1.14 Other 1.06 1.21 0.82 1.32 White 77.27 69.25 73.39 61.91 WOMEN Asian 1.28 2.04 0.35 0.65 Black 0.25 0.40 0.54 1.06 Other 0.15 0.35 0.22 0.69 White 14.16 18.78 22.86 32.03 Total (a) 100.00 100.00 100.00 100.00 Number 53,037 84,045 41,238 65, 170 a/ Excludes those who are retired or not SOURCE: NRC, Survey of Doctorate Recipients. reporting. Women nevertheless appear to be underrepresented in a basic outcome of receiving a doctorate in biomedical or behavioral science: full-time employment as a scientist.24 In 1987 more than 90 percent of the biomedical and 85 percent of the behavioral scientists were so employed; of the remainder, about half were employed part-time (mostly in science); 38 percent were unemployed; and 23.1 percent were unemployed but not seeking a position. Figure 2-16 dramatizes the disproportionately low percentage of eligible women employed full-time in scientific work: 18.4 percent of all women biomedical Ph.D.s in 1987 were not doing full-time science, but this is down from nearly a quarter in 1977. Race, however, was an insignificant factor in determining participation rates in full-time biomedical science. The situation is similar for the behavioral sciences. These data have clear, important implications for science policymakers. If the fraction of biomedical and behavioral science Ph.D.s not doing science full-time remains in excess of IS percent, and if the fraction of degrees awarded to women continues to increase, the projected personnel shortage will be exacerbated (see Chapter 3~. Citizenship Although U.S. science has long relied on the contributions of immigrants, the increasing number of foreign graduate students requires that the role of foreign students and immigration be included in any assessment of the adequacy of the supply of biomedical and behavioral scientists in the United States. In 1988 foreign students earned 18.1 percent of the doctorates in biological sciences awarded by U.S. institutions, up from 11.8 percent in 1980. In the behavioral sciences the proportion earned by foreign students was less: 7.3 percent in 1988:, up from 6.1 percent in 1980. Foreign students who are permanent residents at the time they earn their doctorates 24Another normal outcome, in the biomedical sciences at least, is that of postdoctoral study. Those reporting current postdoctoral study are eliminated from the denominators of percentages reported in Figure 2-16 and the corresponding appendix table. 48
_ 10 ~ ~ 6 ¢- 2 - 0 ~ ~ -6 it? -10 -14 -18 Biomedical Fields _ // . Off . CB" l Behavioral Fields Lo, C] Men ~ Women Asian Black Other White Asian Black Other White SOURCE: Appendix Table A-15. Figure 2-16. Under- and over-representation of race/sex groups in full-time employment in the S&E labor force of biomedical (field percent = 92.3) and behavioral (field percent = 85.6) scientists, 1987. behave very much like U.S. citizens. For example, they virtually all stay in the United States to work.25 During the 1980s, however, about 74 percent of the foreign degree recipients in the biological sciences and about 65 percent in the behavioral sciences were, at graduation, temporary residents who behave very differently from U.S. citizens: only about one-fourth of the temporary residents were still in the United States working or doing postdoctoral study 1-2 years after graduation. Furthermore, evidence shows that foreign nationals who enter the U.S. work force tend to emigrate from the United States at a faster rate than persons who were U.S. citizens at the time of receiving the doctorate. In the life sciences and the social/behavioral sciences, consequently, foreign recipients of U.S. doctorates comprise a comparatively small part of the increases in the doctoral work force--approximately 6 and 4 percent respectively--in contrast to engineering, where they contribute 37 percent of the growth.26 25If we consider all foreign degree recipients, the proportion staying here to work goes up to about 40 percent. See Michael G. Finn, Foreign National Scientists and Engineers in the U.S. Labor Force, 1972-1982, Oak Ridge, TN: Oak Ridge Associated Universities, 1985. 26Michael G. Finn and Sheldon B. Clark, Estimating Emigration of the Foreign-Born Scientists and Engineers in the Unitecl States, Oak Ridge, TN: Oak Ridge Associated Universities, 1988. 49
