Chapter 3:

Science and Engineering Workforce

Labor Market Conditions for Recent S&E Degree-Holders

Bachelor's and Master's Degree Recipients[1]  top

Recent S&E bachelor's and master's degree recipients are a key component of the nation's science and engineering workforce: they account for almost half of the annual inflow to the S&E labor market (NSF 1990, p. 40). The career choices of recent graduates and their entry into the labor market affect the balance between the supply of and demand for scientists and engineers in the United States. Analysis of the workforce status and other characteristics of recent S&E graduates can yield valuable labor market information. This section provides several labor market measures, including median annual salaries and in-field employment rates, that offer useful insights into the overall supply and demand conditions for recent S&E graduates in the United States

Median Annual Salaries top

In 1995, the highest median annual salaries of recent college graduates employed full time were earned by those with engineering degrees. The median annual salary for graduates with a bachelor's degree in engineering was $33,500; it was $44,000 for those with a master's degree. (See appendix table 3-1.) When compared with the salaries for recent science graduates with bachelor's degrees ($22,900) and master's degrees ($35,000), it is apparent that choice of a college major may significantly affect the salaries of recent college graduates entering the labor market.

School Versus Employment top

About one out of four recent S&E bachelor's and master's degree recipients was enrolled in graduate school on a full-time basis in 1995. Students who had majored in the physical and life sciences were more likely to be going on to graduate school as full-time students than were those with degrees in mathematics and the computer sciences or engineering.

In-Field Employment top

Success in the job market varies significantly by level and field of degree. One measure of success is the likelihood of finding employment directly related to a graduate's field of study. S&E master's degree recipients were more likely than bachelor's graduates to find work directly related to their field of study. Approximately one-half of all master's S&E degree recipients—but only a fifth of all S&E bachelor's recipients—were employed in their field of study in 1995. Among both master's and bachelor's degree recipients, students who had received their degrees in either engineering or the computer sciences were more likely to be working in their field of study. Students majoring in the social sciences were less likely to have jobs directly related to their degrees.

Employment Sectors top

The private sector is by far the largest employer of recent bachelor's and master's degree recipients. In 1995, 59 percent of bachelor's degree recipients and 47 percent of master's degree recipients were employed in a private for-profit company. (See appendix table 3-2.) The academic sector is the second largest employer of recent S&E graduates. Master's degree recipients were more likely to be employed in four-year colleges and universities (23 percent) than were bachelor's degree recipients (13 percent). The federal sector employs only 7 percent of S&E master's degree recipients and 4 percent of S&E bachelor's degree recipients. Engineering graduates are more likely than science graduates to find employment in the federal sector. Sectors employing smaller numbers of recent S&E graduates include educational institutions other than four-year colleges and universities, private nonprofit organizations, and state or local government agencies.

Doctoral Degree Recipients   top

Concerns have been raised about labor market opportunities for new Ph.D. scientists and the possible consequences on the health of scientific research in the United States.[2]  Several recent developments have contributed to these concerns, including demographic changes (which have slowed the growth in undergraduate enrollment), reductions in defense and research funding, growth in the importance of Ph.D. programs at foreign schools (see chapter 2, "Worldwide Increase in S&E Educational Capabilities"), and rates of Ph.D. production that approach or exceed the high levels realized at the end of the Vietnam draft.

Since the 1950s, the Federal Government has actively encouraged graduate training in science through a number of mechanisms. However, widespread unemployment or involuntary movement out of S&E by large numbers of new Ph.D. scientists and engineers could have various adverse effects on the health of scientific research in the United States. If labor market difficulties are real but temporary, promising students may be discouraged from pursuing degrees in S&E fields. Eventually, this circumstance could reduce the ability of industry, academia, and government to perform R&D. If labor market difficulties are long term, restructuring will need to take place within graduate education and federal research support to maintain quality research. In either case, when much high-level human capital goes unused, society loses potential opportunities for new knowledge and economic advancement-and individuals feel frustrated in their careers.

Aggregate measures of labor market conditions for recent Ph.D. recipients (one to three years since degree) changed only slightly between April 1993 and April 1995.[3]  The unemployment rate for all recent Ph.D. recipients rose from 1.7 percent in 1993 to 1.9 percent in 1995. (See text table 3-1.) The rate of recent Ph.D.s involuntarily working outside of their degree fields rose slightly, from 4.0 percent in 1993 to 4.3 percent in 1995. These aggregate numbers mask much larger changes in labor market conditions—both positive and negative—within individual disciplines.

Most individuals who complete an S&E doctorate are looking for more than just steady employment at a good salary. Their technical and problem-solving skills make them highly employable, but the opportunity to do the type of work they want and for which they have been trained is important to them. For that reason, no single measure can well describe the S&E labor market. Some of the available labor market indicators are discussed below.

Unemployment Rates top

Only 1.9 percent of recent (one to three years after degree award) Ph.D. recipients were unemployed in April 1995.[4]  (See text table 3-1.) This number is low compared to the 5.7 percent unemployment rate for all civilian workers, and is only slightly higher than the 1.5 percent rate for S&E doctoral recipients. In several fields, however, new Ph.D.s faced higher unemployment rates: 4.3 percent in chemical engineering, 4.0 percent in mathematics, 3.2 percent in sociology/anthropology, and 2.9 percent in physics. While still much lower than for the general population, these unemployment rates are unusually high for a highly skilled group. For recent physics Ph.D.s, however, the 2.9 percent rate represents a large drop from the 5.3 percent unemployment rate reported by the 1993 cohort. On the other hand, the rates for mathematics and chemical engineering are notably greater than the negligible 0.7 and 1.1 percent rates reported respectively in 1993.

Involuntarily Working Outside of Field top

Another 4.3 percent of recent S&E Ph.D. recipients in the labor force reported that they could not find full-time employment "closely related" or "somewhat related" to their degrees. These persons are considered to be IOF—involuntarily out-of-field. This definition of IOF includes those working part time in their fields because full-time work was not available.

As with unemployment, IOF rates varied greatly by field, with 11.2 percent in political science; 9.3 percent in mathematics; 9.1 percent in sociology/anthropology; 6.8 percent in earth, atmospheric, and oceanographic sciences; and 6.7 percent in physics. (See text table 3-1.) Fields with relatively low IOF rates for recent Ph.D.s included 1.0 percent in civil engineering, 2.2 percent for both agricultural and medical sciences, 2.7 percent for both economics and computer sciences, and 2.8 percent in the biological sciences.

Tenure-Track Positions top

Most S&E Ph.D. recipients do not work in academia. (See "How Traditional Is an Academic Career?"; but also see chapter 5, "The Academic Doctoral S&E Workforce.") Across all fields and ages, only 30.8 percent of S&E Ph.D.s in the labor force are in tenure-track or tenured positions at four-year educational institutions. (See text table 3-2.) Across fields, academic tenure-track employment varies from a high of 54.0 percent for economics to a low of 14.0 percent for chemical engineering. Still, the availability of tenure-track positions is an important aspect of the job market for those who do seek academic careers. [Skip Text Box]

How Traditional Is an Academic Career? top

It has long been known from the Survey of Doctorate Recipients (SDR) and other labor force surveys that a large majority of doctorate level scientists and engineers, at any one point in time, work outside academia. The 1995 Scientists and Engineers Statistics Data System (SESTAT) Work History Module, combined with the 1995 SDR core questions, provides current and retrospective career information that allows mapping of typical career paths.

Text table 3-3 divides the population of employed S&E doctorate-holders into four groups: those currently employed as postsecondary teachers, those currently in nonteaching jobs at four-year institutions, those who were formerly postsecondary teachers at some time after completion of their Ph.D.s, and those not currently employed in academia and who reported no postsecondary teaching positions since completion of their Ph.D.s. (Note that tenured administrators and other nonteaching faculty make up most of the difference between the percentage in postsecondary teaching positions and those with tenure or in tenure-track positions; also note that many nonteachers employed in academia also report being former postsecondary teachers.) One weakness of this analysis based on occupation is that it does not capture the past academic affiliations of scientists and engineers who are hired as administrators or researchers without ever being part of the teaching faculty.

A small majority—53.3 percent—of employed S&E doctorate-holders in 1995 were either currently in academia or reported past employment as postsecondary teachers since receiving their degrees. There is less academic involvement in engineering and the physical sciences, where majorities report never having been employed as postsecondary teachers or having no current employment in academia. It is also noteworthy that even in mathematics and the computer sciences, where employment in academia is heaviest, a large majority of currently nonacademic scientists and engineers appears never to have held academic teaching jobs. This view is consistent with shorter career views obtained by longitudinal matching of the SDR data; these data show relatively little movement between academia and industry, excluding new graduates and postdoctorates.

In 1995, 15.9 percent of recent S&E Ph.D. recipients were in tenure-track positions. (See text table 3-2.) This proportion rose to 26.8 percent among those who had received their doctorates within the previous four to six years; it was greater still (30.5 percent) for those at mid-career—11 to 20 years after degree. The percentage of Ph.D.s with tenure-track positions does not, however, reveal much about how difficult it is to obtain academic employment—in fields where many new Ph.D.s prefer employment in industry, there may actually be less competition for academic jobs.

Comparable historical data on tenure-track rates in early career are not available, but comparisons with mid-career tenure-track rates do provide an imperfect indicator of changes in the availability of academic positions. By this relative measure, early career tenure-track rates (four to six years out) are noticeably lower in the biological sciences (-14.4 percentage points), agriculture (-10.1), chemical engineering (-8.6), and physics (-4.7).

The differences in tenure-track rates in the biological sciences are a notable part of a complicated labor market profile for that field. Both unemployment and IOF rates are relatively low in the biological sciences. However, salaries are also lower—and, evidently, so are the opportunities for tenure-track academic employment.

Relationship Between 1995 Occupation and Degree Field top

By a strict definition of occupational titles, 31.5 percent of employed recent Ph.D.s were in occupations outside science and engineering, often with administrative or management functions. When asked how related their jobs were to their highest degree, only a small proportion of recent Ph.D.s in non-S&E occupations said that their jobs were unrelated to their degree. (See text table 3-4.) By field, these respondents ranged from 1.5 percent of recent engineering Ph.D. graduates to 4.5 percent of recent Ph.D. graduates in mathematics and the computer sciences.

Changes in Employment Status top

Of the 72.2 percent of recent S&E Ph.D. recipients who were in "regular" employment in 1993 (that is, not in a postdoctorate appointment and not involuntarily working outside of their fields), the vast majority—94 percent—were still in regular employment in 1995. (See figure 3-1.) Of those in other 1993 employment statuses (postdoctorate, IOF, or unemployed), 50 percent of each group had moved to regular employment by 1995. Forty-five percent of 1993 postdoctorates were still in a postdoctorate position in 1995; 37 percent of those working involuntarily outside of their fields were IOF in 1995 as well. There was, however, much less evidence of long-term unemployment: only 0.3 percent were unemployed in both 1993 and 1995.

Median Annual Salaries top

The median salary earned by recent science and engineering Ph.D. recipients in 1995 was $40,000, with the highest median found for engineering Ph.D.s ($54,000) and the lowest for Ph.D.s in the life sciences ($32,000). Despite the wide variety of employment types and fields for new Ph.D. recipients, there is a fairly narrow distribution of salaries around this median—the 10th percentile makes $22,500 and the 90th percentile, $65,000. (See text table 3-5.) The lowest 10th percentile salary ($8,000) is found for recent Ph.D. recipients in sociology/anthropology. The highest 90th percentile salary was $85,000, for recent Ph.D. recipients in the computer sciences and economics.

Another meaningful way to view new doctorate salaries is by sector of employment. Median salaries in 1995 for recent Ph.D.s were highest in the private, noneducation sector ($56,000) and lowest for postdoctorates ($28,000). (See text table 3-6.) Government salaries tended to be just slightly above those of tenure-track positions in academia.[5]  While the pattern of salary by degree field also varied by sector of employment, salaries were generally higher in engineering and math/computer sciences and lower in the social and life sciences.

Postdoctorate Appointments top

A postdoctoral appointment is defined here as a temporary position awarded in academia, industry, or government primarily for the purpose of gaining additional training in research. This definition is used in the 1995 Survey of Doctorate Recipients to ask respondents about current and past postdoctorate positions they have held.[6] 

Data and analyses on postdoctorates are closely related to recent Ph.D. labor market issues. In addition to gaining more training, recent Ph.D. recipients may accept a temporary, usually lower paying, postdoctorate position because a more permanent job in their field is not available. NSB (1996) reported that there was no strong evidence that the number or length of postdoctorate positions was being driven by changes in labor market conditions. With the new data provided by an extensive postdoctorate module in the 1995 Survey of Doctorate Recipients, some labor market effects can now be discerned in some specific disciplines.

Reasons for Taking a Postdoctorate. The most commonly reported reason given by 1995 postdoctorates for taking a postdoctorate appointment was to acquire additional training in their Ph.D. field (35.4 percent).[7]  Other respondents reported that they were taking a postdoctorate to receive training outside of their respective Ph.D. field[8]  (18.5 percent) or to work with a particular researcher or institution (21.5 percent). Text table 3-7 shows reported reasons for taking a postdoctorate in the six fields that accounted for 92 percent of 1995 S&E postdoctorate appointments.

Beyond these traditional uses of a postdoctorate, 17.1 percent of respondents reported that they accepted a postdoctorate appointment because other employment was not available. This proportion rises to 29.3 percent in the earth, atmospheric, and oceanographic sciences and to 26.8 percent in physics-two fields with relatively high unemployment and IOF rates among recent Ph.D. graduates.

Incidence and Length of Postdoctorate Appointments. Although there are some postdoctorate positions in all academic disciplines, most are concentrated in a small number of fields in which postdoctorate appointments are part of a traditional career path. Although some scientists and engineers appear to take postdoctorate positions at all points in their careers, they usually do so within a few years of completing their doctorate. (See figure 3-2.) The incidence of postdoctorate appointments is greatest in the biological sciences and physics, but few are in postdoctorate positions in these fields beyond six years after degree award.[9] 

Text table 3-8 provides information from the SDR Postdoctorate Module on the proportion of each graduation cohort that ever held a postdoctorate position and the median number of months in postdoctorates for those who held them.[10]  For the more recent cohorts, particularly those only one to three years since degree, length of time in postdoctorate is constrained by the limited time they have held their degrees.

Across all fields, the Postdoctorate Module shows a steady increase over time in both the incidence and length of postdoctorate experiences. It is difficult to tie these trends directly to labor market events or even to claim a consistent pattern across fields. In physics, chemistry, and psychology—fields with distinct labor markets—the incidence of postdoctorates between one to three and four to six years after degree has risen, despite the lesser opportunity of the younger cohort to obtain a postdoctorate. In psychology, the agricultural sciences, and chemistry, there is no trend toward longer postdoctorate appointments. In the biological sciences, even the mid-career cohort—11 to 20 years after degree—had a very high incidence (66.3 percent) and length (38 months) of postdoctorate positions.

Postdoctorate Transitions: What Were 1993 Postdoctorates Doing in 1995? Of those in postdoctorate positions in April 1993, 41.6 percent were still in a postdoctorate position in April 1995. (See text table 3-9.) Only 12.1 percent transitioned from a postdoctorate to a tenure-track position at a four-year educational institution; 21.2 percent found other positions at educational institutions, and 16.6 percent went to work for a private for-profit firm.

The percentage of postdoctorates obtaining tenure-track positions is not large even for those with greater time since degree—only 18.8 percent of 1993 postdoctorates who were five to six years since degree were in tenure-track positions in 1995. (See text table 3-10.) This is, however, a much greater rate of transition to permanent academic jobs than for postdoctorates one to two years since degree (10.4 percent). One in five is still a low rate if an academic career is viewed as the primary objective of most Ph.D. scientists accepting a postdoctorate appointment at that point in their career.

For those in postdoctorates seven or more years after their degree, the rate of transition to tenure-track appointments drops to 9.8 percent. To a great extent, this rate is driven by career patterns in the biological sciences, where there have long been large numbers of Ph.D. scientists pursuing multiple postdoctorate appointments. However, in physics—where multiple postdoctorates are a more recent phenomenon[11] —the percentage of postdoctorates transitioning to tenure-track appointments begins to drop much earlier (three to four years since degree), to 7.1 percent.

For both physics and the biological sciences, the unemployment rate in 1995 for 1993 postdoctorates was greatest for those with more time since degree-3.4 percent for biological scientists seven or more years since degree and for physicists three to four years after degree. There was also an increase in the rate of transition to the "other education" category. This category includes some individuals who become adjunct faculty, but it primarily encompasses other non-tenure-track research and administrative jobs at a university.

green bar


[1] Data in this section are taken from the 1995 National Survey of Recent College Graduates. This survey collected information on the 1995 workforce/other status of 1993 and 1994 bachelor's and master's degree recipients in S&E fields. Surveys of recent S&E graduates have been conducted biennially for the National Science Foundation since 1978. For information on standard errors associated with survey data, see NSF (1997b).

[2] For a more detailed discussion, see COSEPUP (1995).

[3] This section primarily uses data from the 1993 and 1995 Survey of Doctorate Recipients (SDR), a biennial National Science Foundation (NSF) survey of doctorate-holders from U.S. institutions up to age 75; and the closely related Scientists and Engineers Statistics Data System (SESTAT) integrated file which contains data from the SDR and two other NSF surveys, the National Survey of College Graduates and the National Survey of Recent College Graduates. For more information on SDR, see chapter 5, "Data Sources: Nature, Problems, and Comparability."

[4] People are said to be unemployed if they were not employed during the week of April 15, 1995, and had either looked for work during the preceding four weeks or were on layoff from a job.

[5] Salaries reported on an "academic year" basis have not been adjusted upwards, as was done in pre-1996 volumes of Science & Engineering Indicators.

[6] It is clear, however, that the exact use of the term "postdoctorate" differs among academic disciplines, among different universities, and among the different sectors that employ postdoctorates. It is likely that these differences in labeling affected self-reporting of postdoctorate status on the Survey of Doctorate Recipients.

[7] A recent joint National Science Foundation-French National Center for Scientific Research (CNRS) project to study French doctorates and postdoctorates in the United States showed a similar pattern. Although not a fully representative sample, many of the respondents noted that the reason they took a postdoctorate in the United States was to improve their job opportunities in France (see Terouanne 1997).

[8] Many respondents to this question may have interpreted "field" very narrowly, so training outside of their field may simply refer to a subfield of their discipline that lies outside their dissertation work.

[9] The profile of those who had a postdoctorate in 1995 does not reveal much about the length of time spent in postdoctoral appointments—a person in a postdoctorate six years after obtaining a Ph.D. may have just begun the appointment. The profile also does not reveal much about how postdoctorates today differ from their historical patterns.

[10] Recall bias may well lead to underreporting of postdoctorate experiences by older cohorts, but this occurrence may be less problematic than comparisons of reported postdoctorate rates among the sometimes dissimilar survey instruments used over the years. For length of appointment, up to three postdoctorates reported in the Postdoctorate Module are aggregated.

[11] See text table 3-8 for the historical pattern of postdoctorates. Due to the small numbers of physicists in postdoctorates beyond four years after their degree, there was not a sufficient sample size to estimate transition rates.

Previous | Top | Next

Indicators Home | Contents | Help | Comments