May 1998

In recent years, a great deal of concern has been raised regarding employment prospects for PhD recipients. Although, historically, PhDs have had the lowest unemployment rates of any group in the U.S. workforce, several professional societies in the sciences have reported double-digit PhD unemployment rates. The 1995 Massy-Goldman study of PhD production projected unemployment rates as high as 50%.¹ This report was cited by The New York Times and was incorrectly referred to as a "survey" rather than a simulation. Many articles followed in the popular press, both nationally and locally, all of which projected dismal employment prospects for PhD recipients and voiced concern regarding "over production" of doctoral degrees. Some of these reports also stated that increasing numbers of doctoral degree holders would be employed in positions unrelated to the PhD or locked in temporary positions.

While much of this concern is well placed, there is a great deal of disagreement regarding the validity of the data. After the Massy Goldman model was published, flaws were discovered which indicated a much lower unemployment rate than that projected. A study conducted by the National Science Foundation (NSF) reported an unemployment rate of just 1.6% and an underemployment rate of 4.3%.² It is interesting to note that the NSF study reporting low unemployment rates covered the same period of time in which several science professional societies reported double-digit unemployment.³ The discrepancy between the reports may reflect the timing of the surveys. The NSF queried degree recipients nearly a year after graduation, while the professional societies contacted PhDs within six months of graduation, i.e. before positions may have been secured.

In addition to disagreements over unemployment rates, there is also debate about the extent to which PhDs are "underemployed."⁴ A nation-wide increase in production of PhDs has not been matched with a corresponding increase in the number of available tenure-track faculty positions. As a consequence, PhDs are increasingly finding employment in business, governmental and other "non-traditional" settings. While some would contend that such employment constitutes "underemployment" others see it as a positive development in which new and vital roles are emerging for PhDs outside of academe. A more concrete source of concern about underemployment can be found in reports showing that increasing numbers of PhDs are employed in part-time or temporary positions, or in positions unrelated to their doctoral degree. There is also evidence that PhD recipients are increasingly employed in "postdoc" positions, and are remaining in such positions for longer periods of time before securing permanent employment. According to NSF, there were 26,000 science and engineering postdocs in doctorate-granting institutions in 1995, which represents a 32% increase over the number of postdocs in 1988. The NSF study showed that 39% of all biological science PhDs were in a postdoc 3-4 years after degree, and that 18% were still in postdoc status 5-6 years after degree.⁵

The number of doctoral degrees granted by US academic institutions has increased 28% in the past ten years.⁶ International students accounted for much of this increase. The number of PhDs granted to international students increased by 86% over the 10-year period, while the number granted to US citizens increased by only 21%. The Committee on Science, Engineering, and Public Policy (COSEPUP) reported that international PhDs made up one third of graduate students in 1992, up from one fourth a decade earlier.⁷ This increase may be partly due to a 1992 change in immigration law that raised occupational quotas. According to the NSF, the number of postdocs held by foreign nationals in 1995 exceeded the number of postdocs held by US citizens or legal immigrants.⁸

From 1987 to 1996, UW doctoral degree production increased 34% (Appendix A.1). During that period, international student PhD production at the UW did not show a consistent upward or downward trend (Appendix A.2).

Both nationally and at the UW, there has been a significant increase in the number of PhDs granted to women and minorities over the past ten years. Nationally, the number of doctoral degrees granted to women increased 54% from 1987 to 1996. The rate of increase at the UW, at 62%, is even higher during this period (Appendix A.3). Minority PhDs increased 122% nationally from 1987-1996. This compares with an increase of 133% at the UW during the same period (Appendix A.4).

The increase in doctoral degree production has been sharper in some disciplines than in others, both nationally and at the UW. By discipline, the largest national increases over the ten year period were in engineering (70%), humanities and arts (49%), and natural sciences (39%). In 1996, the largest share of degrees granted nationally were in the natural sciences (35%). At the UW, a similar increase in PhD production was seen in engineering (64%), and an even larger increase in humanities fields (121%). The increase in the natural science fields at the UW was more modest (19%) (Appendix A.5). At the institutional level, such changes need to be examined within context, as a year with an unusually low (or high) number of graduates can artificially skew the apparent percentage change.

Recent reports indicate that doctoral degree production is beginning to slow. The National Science Foundation reported a modest 1.6% increase from 1995 to 1996. That reports states that "while the number of PhDs awarded has increased since 1986, the rate of growth in the past decade has not matched the rate of growth in the 1960s and has generally been below the average annual growth rate of 4 percent for the past 40 years."⁹

In response to the issues noted above, the Graduate School conducted a survey during the 1996-97 academic year. The survey posed questions regarding the current employment status of 4,369 persons who received doctoral degrees from the UW over the period from Summer Quarter 1986 through Spring Quarter 1996. Supervisory committee chairs (i.e. the students’ major professors) were asked to complete the questionnaire on behalf of their doctoral recipients. A total of 3,834 forms were returned (87.7% response rate) of which 3,500 contained known employment data (80.1%). This high response rate would not have been possible with the logistical difficulties of direct mailings to the PhD recipients themselves.

The survey instrument was kept simple to reduce the possibility of ambiguous responses (Appendix C). Only five questions were asked:

1. What is the PhD recipient’s current primary employment?
2. Where is the position located?
3. Is it a part-time position?
4. Is it a temporary position?
5. Is the position unrelated to the PhD?

The questionnaire also contained an optional Comments section. Many respondents used this section of the form to provide detailed information regarding their students’ current employment.

Response rate by School/College is shown in Appendix A.6; by year of graduation in Appendix A.7.

Among the 4,369 students who received PhD degrees during the ten year period covered by the study, 37% were women, 6% were minority, and 88% were U.S. citizens. At the time of graduation, 18% of the new PhDs were under 30 years of age, 38% were ages 30-34, and 22% were ages 35-39. Twenty-one percent were over 40 at the time of degree completion. The average age of those receiving degrees in 1986-87 was 34.5, for those receiving degrees in 1995-96 it was 35.5. Over half of the PhDs (54%) received degrees in the natural sciences and engineering.

As noted above, current employment data was obtained for 3,500 PhD recipients. The demographic characteristics of that subset were representative of the entire group.

It has been widely reported that PhD recipients face increasing competition for "traditional" positions as university faculty. This is due, reports suggest, to continued erosion of university budgets and a decline in the rate of faculty retirements. The National Center for Educational Statistics reports that "Of the more than 500,000 full-time instructional faculty, only 7% said they were likely to retire in the next three years."¹⁰ According to the COSEPUP report, over 50% of 1969-72 PhD recipients were employed in academe 5-8 years post graduation, compared with only 45% of 1983-86 graduates. The COSEPUP report stated that, "After 5-8 years, 26% of science and engineering PhDs were employed in business and industry as of 1973, a proportion that grew to about 45% in 1991."

The UW survey findings were largely in agreement with those of the COSEPUP report. Thirty-seven percent of the degree recipients were reported to be employed in tenure-track positions (Appendix B). It is not surprising that the UW, a "research 1"¹¹ institution, has a somewhat higher percentage of graduates in tenure-track positions than the 30% indicated in the COSEPUP report, which includes all doctoral institutions. Men and women held tenure-track positions in similar proportions, but women were more likely than men to be in non-tenure-track faculty positions and less likely than men to be employed in non-faculty positions (Appendix A.8). A greater proportion of international students (45.3%) were in tenure-track positions than were U.S. citizens, with only 29.5% of minority and 36% of non-minority PhDs holding such jobs (Appendix A.9).

There was disparity among disciplines with respect to the percentage of PhDs reported to be in tenure-track faculty positions (data not shown). More than 80% of Business School PhDs were in tenure-track positions and 63% of humanities graduates. By contrast only 19% of graduates from the Schools of Medicine and Pharmacy held tenure-track positions. Thirty-nine percent of the School of Medicine graduates held postdoctoral appointments.

The UW PhD survey is of limited utility for gauging employment trends, since graduates were surveyed at a single point in time. However, the survey data did provide a comparison of early-career versus mid-career PhDs. Not surprisingly, early-career graduates held fewer tenure-track positions (Appendix A.10) and more postdocs (Appendix A.11) than did the mid-career group.

Doctoral degree granting universities were the major employer of UW graduates, with 43% of the PhDs employed at such institutions (Appendix B). Fifty-eight percent of those at doctoral institutions were employed in tenure-track positions. A somewhat larger share of the international PhDs (50%), and a somewhat smaller share of minority PhDs (35%) were employed at doctoral institutions (Appendix A.13). Thirteen percent of the graduates were reported to be employed at non-doctoral degree granting baccalaureate institutions and 2.4% at community colleges (Appendix B).

The School of Social Work had the highest proportion of graduates (73%) reported to be employed at doctoral institutions, followed by the School of Law (67%), Public Health and Community Medicine (58%), and School of Medicine (55%) (Appendix A.16). Employment at non-doctoral institutions was highest for graduates of the Arts and Sciences – Humanities (36%), Business Administration (34%), and the School of Nursing (29%).

Eighteen percent of the graduates were reported to be employed in business/industrial positions. Nine percent of women held positions in business/industry, 24% of international students, and 21% of minorities.

The largest percentages of employment in business and industry were reported among graduates of the School of Pharmacy (69%) and the College of Engineering (50%).

Forty one percent of all US citizen graduates were reported to be employed in Washington State (Appendix A.17). Of those with business/industrial employment, 52% were employed in Washington State (Appendix A.18). A somewhat higher percentage of women (45%) than men (39%) held positions in Washington State.

Only 37% of graduates from the first three years (1986-88) of the survey period were employed in Washington State, but 47% of the 1996 cohort were employed in state (Appendix A.19c). The basis of this difference is not known, however, it is not due to a concentration of postdoc positions in Washington. It is possible that some degree recipients begin their careers in Washington and leave the state at a later date. Unfortunately, the data do not permit that type of analysis.

Out of 3,500 PhD recipients whose current employment status was known, only 43 (1.2%) were reported to be unemployed. This correlates closely with the science and engineering unemployment rate of 1.5% reported by the NSF in 1995 (Appendix A.20)¹². When the most recent degree recipients are excluded and only US citizens are considered, fewer than 1% of the UW graduates are unemployed. As would be expected, this is far below the unemployment rate for the general populace. Unemployment rates by School/College are listed in Appendix A.21.

While only 37% of the total cohort were women, nearly three-quarters of persons reported to be unemployed were women (Appendix A.23). This notable gender gap was relatively constant throughout the ten year cohort: 78% of the 1995-96 unemployed graduates, 71% of the 1994-95 unemployed graduates, and 73% of 1986-93 unemployed graduates were women. Nearly 90% of the unemployed were US citizens.

The unemployment rate varied among disciplines. Although the University-wide average was similar to the low unemployment rate reported by NSF, unemployment was somewhat higher in disciplines that have reported high unemployment nationally. Among the 43 individuals who were reported to be unemployed, 10 received degrees in natural science fields; however, the proportion of unemployed natural science graduates was lower (under 1%) than for all other fields. The fields reporting the highest percentages of unemployment were Bioengineering, Germanics, Materials Science & Engineering, and Special Individual PhD, each with an employment rate exceeding 5% (Appendix A.21). UW humanities unemployment, at 1.5%, is less than half that reported by the National Science Foundation in 1995.

Unfortunately, the survey did not ask whether the unemployed were actively seeking work. Some clearly were not in the job market, as indicated by a sample of comments below:

On the other hand, some comments portray unsuccessful attempts to secure employment and echo anecdotes that are frequently cited in national studies:

Part-Time Employment--In the UW survey, nearly 4% of the positions were classified as being part-time. While only 2% of international PhDs, and 3% of minorities worked part-time, 6% of women PhDs were reported to be employed part-time. NSF similarly found that female science and engineering PhDs were much more likely than their male counterparts to be employed part-time.¹³

As one would suspect, the early-career cohort were more likely to be employed part-time (5%), but a significant number of mid-career PhDs were also reported to be in part-time positions. The greatest concentration of persons holding part-time positions were from humanities and arts fields (8%); the fewest were in engineering and education (1%). When considering only 1995-96 degree recipients, 16% of social science graduates were reported to be employed part-time and 11% of humanities and arts graduates.

Often, survey comments indicated that part-time employment was by choice:

However, other comments, indicated that part-time employment was not a desired goal:

Temporary Employment--Five percent of the graduates were reported to hold temporary employment. International students and minorities reflected a somewhat lower percentage at 3% of their respective totals. Nearly half of the temporary positions were described as postdoc appointments. Comments for those holding temporary positions that were not postdocs generally reflected job market frustration:

Employment Unrelated to Field of PhD Study--Eighty-three graduates, making up 2% of total survey cohort, were reported to be employed in jobs that were unrelated to their PhD. When broken down demographically, 2% of females, 3% of males, 3% of minorities, and 2% of international students were reported to be employed in fields unrelated to their PhD.

Overall Underemployment--A combination of all graduates reported to be employed in part-time work, temporary work, and work unrelated to the PhD made up 10% of the total: females were the most likely to be doing this type of nontraditional work (15%), males were less likely (7%), as were international students ( 7%) and minorities (9%).

Underemployment rates (including part-time, temporary work, and work unrelated to the Ph.D.) by School and College are shown in Appendix A.24.

Results of the UW survey confirm observations that have been made nationally about the changing nature of postdoctoral employment. The survey asked respondents to identify whether their student’s current employment was a "postdoc" and how many postdocs had been held. For this analysis, the graduates were divided into two groups; an early-career cohort (those completing their Ph.D. in the past four years) and a mid-career cohort (those completing their degrees within five to ten years of the survey).

Fifty-five percent of the surveyed graduates reportedly held at least one postdoctoral appointment and this proportion was constant across the ten years. Fourteen percent were reported to have held two postdocs and 4% three or more (two individuals had six postdocs).

Seventy eight percent of those with degrees from the School of Medicine and 68% of those with degrees from Arts and Sciences – Sciences were reported to have held one or more postdocs. Surprisingly 72% of Arts and Sciences – Arts graduates were reported to have held one or more postdoc and 76% of those in Arts and Sciences – Humanities. The latter reports (i.e. Arts and Humanities) appear to include temporary teaching appointments under the category of "postdoc" (Appendix A.26).

Eighteen percent of recent graduates were reported to be in postdoctoral positions at the time of the survey. Women, minorities and non-citizens were represented proportionately.

Most postdocs (72%) were at doctoral degree granting institutions. Only 1% were at colleges and universities which do not grant the PhD. Nineteen percent were at national laboratories, government agencies, or business and industry.

Early-career postdoctoral employment was most prevalent among those receiving degrees from the School of Medicine (65%), and Arts and Sciences – Sciences (32%). Few postdocs were reported for graduates in the social sciences, humanities, and arts.

It is somewhat surprising that 84 mid-career graduates were reported to still be in postdocs at the time of the survey. Although this represents only 4% of the mid-career cohort, postdoc employment after six years lies somewhat outside the traditional definition of a postdoc. A closer look at this group of 84 individuals reveals that 33 received degrees in Arts and Sciences – Sciences and 19 received degrees from the School of Medicine (Appendix A.27). Extended postdocs represent 7% of the total Arts and Sciences—Sciences mid-career cohort and 16% of the School of Medicine mid-career cohort. National statistics show that 18% of biological science graduates are still in postdoc status 5-6 years after degree completion.¹⁴ In the UW survey the highest proportions of extended postdocs were reported in Microbiology (32%), Astronomy (21%) and Genetics (18%).

The UW survey revealed interesting information regarding the number of postdocs held by an individual. Five percent (N=101) of the mid-career cohort had held three or more postdocs (Appendix A.28). Nine percent of Arts and Science – Science graduates had held three or more postdocs. It is interesting to note that some of the disciplines for which national data portray difficult times for PhD employment also reported the highest percentage of three or more postdocs in the UW survey. Thirty percent of Mathematics PhDs were reported to have three or more postdocs and 25% of Astronomy PhDs. In the social sciences, 21% of Anthropology graduates had three or more postdocs (Appendix A.29).

Mark Regets (NSF SRS Division) has noted that, "Anecdotally, one motive for accepting multiple postdocs is to enhance one’s credentials for academic employment. If this is true for many of those in multiple postdocs, it does not appear to be a successful strategy. Of those in first postdocs in April 1993, 13% were in tenure-track positions at 4-year institutions in April 1995. For those in their second or greater postdoc, the percentage transitioning to tenure-track employment was only slightly higher – 16.2%."¹⁵

The UW survey confirms this observation. Thirty-eight percent of mid-career graduates who had held no postdocs, and 50% of those with one postdoc were reported to be in tenure-track positions. By contrast, only 19% of those with two or more postdocs were employed in tenure-track positions.

Noting the growing trend among PhDs to find employment in nonacademic settings, the COSEPUP report recommended that "the graduate education enterprise – particularly at the department level – implement several basic reforms to enhance the education experience of future scientists and engineers who will work in either academic or nonacademic settings. If programs offer a wider variety of degree and curricular options that are valued by their faculty, students will be better served. In addition, we have an obligation to inform graduate students accurately and explicitly about career options so that they will be able to make better educational choices, formulate more realistic career expectations, and achieve greater satisfaction in their careers while contributing more effectively to fulfilling national goals."

Although some have called for limits on PhD production in certain fields, others have cautioned against such measures in reaction to employment trends that are, as yet, not well understood. According to the Report of the Ad Hoc Working Group on the Supply of Science, Engineering and Mathematics (SEM) Professionals, "It is not currently possible and will probably never be possible to predict with a high degree of accuracy shortages or surpluses of sciences and engineers several years into the future."¹⁶ Geoff Davis, Assistant Professor of Mathematics at Dartmouth, argues that the history of mathematics PhD production has been one of continual alteration between shortage and surplus conditions and that this oscillation is an inherent function of the PhD production process. "Market conditions have a powerful influence on first-year graduate enrollment levels. When demand for doctorates is high, enrollment levels surge, and when demand is low, they fall. Because of the lag between enrollment and graduation, however, this feedback results in a delayed response to market conditions and leads to oscillations in levels of PhD supply."¹⁷ He goes on to say that the key to damping the oscillations is to use future market conditions in determining first-year enrollments. The problem in this, he says is "the present system of incentives rewards only one thing: research productivity. Institutions have no stake in the career outcomes of their graduates. We neither reward nor recognize programs that are especially successful at placing their students. As a result, programs lack strong motivation for change. Moreover, because such reforms as broadening curricula can take student time away from research, these changes actually penalize programs." He concludes by recommending that program outcome data be made available to applicants.

In the article Doctoral Education: The Short-Term Crisis Vs. Long-Term Challenge,¹⁸ Roger Geiger recommends a number of "structural adjustments" to doctoral education including, "finding additional clientele whom doctoral education might serve, how to minimize economic disincentives and how to work with, rather than against, the quality imperative and the prestige hierarchy." To broaden the focus beyond teaching and research, he asserts that the doctoral degree "will have to become, in some fields, a valued credential in the private sector." Regarding the "horrendous economics of doctoral education", he suggests "the only way to diminish these disincentives, and to fashion economically viable doctoral degrees, is to provide a more streamlined version of the doctorate—in other words, to shorten the time-to-degree." To work with, rather than against, the quality imperative and the prestige hierarchy, he suggests "the doctoral process ought to be separated in to what are now in fact three distinct phases: the pre-dissertation stage, the dissertation stage, and the postdoctoral stage…Completion of the pre-dissertation work should represent the end of a definite stage of training and a decision point about future careers… A decision after the pre-dissertation stage could promote interdisciplinary competence by encouraging students to move to adjacent fields…The decision point at the transition from Ph.D. to postdoctorate, under current circumstances, probably extends already lengthy times-to-degree. Because of stiff competition for the most attractive postdoctoral appointments, students prolong dissertation research to strengthen their credentials. Because students at this phase are productive research assistants, contributing to the output of laboratories, such extensions dovetail with the interests of their mentors. The process also harmonizes with the qualitative imperative, thus encouraging even higher, more time consuming, attainments. This cycle cannot be easily broken in the absence of greater demand to lure doctoral candidates from their laboratories. The frank recognition that the definitive competition for faculty positions takes place at the postdoctoral level (Zumeta, 1985) is nevertheless the prerequisite for addressing this problem. Given this reality, academic departments ought to be more sympathetic toward streamlining doctoral programs, optimizing requirements, and circumscribing the extent of the Ph.D. thesis."

The UW survey has provided valuable insights, but has highlighted the need for additional data. In particular, trend data is largely absent. Also, a few additional questions (e.g. is the person actively seeking employment?) would increase the value of the data. To address these issues, the PhD Employment Survey form will be revised and a new survey will be conducted at a date to be determined. In addition, the survey instrument issued by the UW Educational Assessment Office is being revised to address questions relating specifically to PhD employment. Since that survey is mailed directly to graduates about eleven months after graduation, it will provide an effective complement to the data reported here.