Recent S&E bachelorís and masterís degree recipients form a key component of the Nationís S&E workforce; they account for almost half the annual inflow to the S&E labor market. 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 that offer useful insights into the overall supply and demand conditions for recent S&E graduates in the United States. Among these measures are median annual salaries, unemployment rates, and in-field employment rates.
In 1997, the highest median annual salaries for recent full-time employed graduates with bachelorís degrees in the sciences went to those with degrees in computer and information sciences ($37,700), and the highest salaries among those with degrees in engineering went to those with degrees in electrical, electronics, computer, and communications engineering ($40,500). (See appendix table 3-18.)
The same pattern was true among recent graduates with masterís degrees. The highest median annual salaries went to graduates with masterís degrees in computer and information sciences ($51,200) and electronics, computer, and communications engineering ($55,000).
Approximately one-fifth of 1995 and 1996 bachelorís and masterís graduates were enrolled in graduate school on a full-time basis in 1997. Students who had majored in the physical and related sciences and the life and related sciences were more likely to be in graduate school as full-time students than were graduates with degrees in computer and information sciences or engineering. (See appendix table 3-18.)
Although individuals use college degrees to enter a wide variety of career paths, the extent to which their employment is related to their degrees may be one indicator of the vocational relevance of a degree. Across all fields of S&E in 1997, 70.4 percent of recent bachelorís degree graduates and 91.4 percent of recent masterís degree graduates said their jobs were related to their field of degree (appendix tables 3-3 and 3-4). At the bachelorís level, employment related to field of degree for recent S&E graduates varied from 58.8 percent in the social sciences to 92.9 percent in computer sciences and 89.3 percent in engineering. At the masterís degree level, there is much less variation by field of degreeranging from 87.6 percent of recent masterís degree graduates in social sciences saying their jobs are related to their degrees, to 97.9 percent of recent computer sciences masterís degree graduates.
The private, for-profit sector is by far the largest employer of recent bachelorís and masterís S&E degree recipients. (See text table 3-5.) In 1997, 66 percent of bachelorís degree recipients and 59 percent of masterís degree recipients were employed in private, for-profit companies. The academic sector has been the second largest employer of recent S&E graduates. Masterís degree recipients were more likely to be employed in four-year colleges and universities (9 percent) than were bachelorís degree recipients (5 percent). The Federal sector employed only 7 percent of S&E masterís degree recipients and 4 percent of S&E bachelorís degree recipients in 1997. Engineering graduates are more likely to find employment in the Federal sector than science graduates. Other sectors employing small numbers of recent S&E graduates include educational institutions other than four-year colleges and universities, nonprofit organizations, and state and local government agencies.
Concerns about the labor market for workers with doctorates in S&E often focus on recent Ph.D. recipients entering the labor market and attempting to begin a career. Although the vast majority of S&E Ph.D. recipients find work that is related to their degrees, there is concern that fewer opportunities may make doctoral level science careers less desirable.
Since the 1950s, the Federal Government has actively encouraged graduate training in science through a number of mechanisms. Real or perceived labor market difficulties for new Ph.D. scientists and engineers, however, 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, transfer knowledge, or perform many of the other functions of scientists in the modern economy. If labor market difficulties are long term, restructuring may need to take place within graduate education both to maintain high-quality research and to prepare students better for a wider range of career options. In either case, when much high-level human capital goes unused, society loses opportunities for new knowledge and economic advancement, and individuals feel frustrated in their careers.
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 describe well the S&E labor market. Some of the available labor market indicators are discussed below.
Aggregate measures of labor market conditions changed only slightly for recent doctorate recipients in S&E (defined here as those one to three years after their degree). Unemployment fell from 1.9 percent for a similar graduation cohort in 1995 to 1.5 percent in 1997. (See text table 3-6.) At the same time, the proportion of recent Ph.D. recipients reporting that they were either working outside their field because a job in their field was not available, or that they were involuntarily working part-time, rose slightly from 4.3 percent to 4.5 percent. These aggregate numbers mask a number of changesboth positive and negativein a number of individual disciplines. In addition, in many fields the involuntarily out of field (IOF) and unemployment rates moved in opposite directions. In many ways, whether highly skilled individuals who are unable to get the type of employment they desire become unemployed or accept employment outside their field, may reflect the type of expectations they have of the labor market.
Even compared to relatively good labor market conditions in the general economy, the 1.5 percent unemployment rate for recent S&E Ph.D. recipients is very lowthe April 1997 unemployment rate for all civilian workers was 5.0 percent. (See the sidebar, "Data on Recent Ph.D. Recipients in Professional Society Data.") In 1995, recent graduates in several Ph.D. disciplines had unemployment rates above 3 percentstill low, but unusually high for a highly skilled group. Between 1995 and 1997, unemployment rates fell for recent Ph.D. recipients in most disciplines, but increased in a few. The largest increase was in chemistry, where the unemployment rate for recent Ph.D. recipients rose from 2.2 to 3.5 percentalso making chemistry the field with the greatest unemployment rate for recent Ph.D. recipients. In 1997 unemployment rates of less than 1 percent were found for recent Ph.D. recipients in civil engineering (0.0 percent), mechanical engineering (0.5 percent), electrical engineering (0.6 percent), mathematical sciences (0.6 percent), computer sciences (0.7 percent), physics-astronomy (0.7 percent), and economics (0.9 percent).
In 1998, data from surveys of new Ph.D. recipients for 1996Ė97 conducted by 13 S&E professional societies on 14 different disciplines were coordinated by the Commission on Professionals in Science and Technology. A common set of core questions was used in each societyís poll of its own doctoral graduates to allow for collection of comparable career-related data. One of these common data elements, the unemployment rate is shown in text table 3-7. Unemployment ranged from 1.8 percent for recent physics Ph.D. recipients to 7.0 percent for recent Ph.D. recipients in political science.
Another 4.5 percent of recent S&E Ph.D. recipients in the labor force reported that they could not find full-time employment (if they wished full-time employment) that was "closely related" or "somewhat related" to their degrees. Although this is a more subjective measure than unemployment rates, it often provides a more sensitive indicator of labor market difficulties for a highly educated and employable population. It is best used, however, along with the unemployment rate as measures of two different forms of labor market distress.
The highest IOF rates in 1997 were found for recent Ph.D. graduates in physics-astronomy (12.2 percent) and in mathematical sciences (11.0 percent). These two fields also had among the lowest unemployment rates, although in physics-astronomy the increase in the IOF rate from 1995 to 1997 was much greater than the decrease in unemployment. The lowest IOF rates were found in computer sciences (2.1 percent) and the biological sciences (2.2 percent).
Most S&E Ph.D. recipients do not ultimately work in academiain most S&E fields this has been true for several decades. (Also see chapter 6, "Academic Research and Development: Financial and Personnel Resources, Support for Graduate Education, and Outputs.") In 1997, of S&E Ph.D. recipients four to six years after receipt of their degrees, only 22.9 percent were in tenure track or tenured positions at four-year institutions of higher education. (See text table 3-8.) Across fields, tenure-program academic employment four to six years after Ph.D. ranged from 11.9 percent in chemical engineering to 51.2 percent in sociology-anthropology. For Ph.D. recipients one to three years after their degrees, only 16.0 percent were in tenure programs, but this reflects the increasing use of postdoctoral appointments (or postdocs) by recent Ph.D. recipients in many fields.
Although academia must be considered just one possible sector of employment for S&E Ph.D. recipients, the availability of tenure-track positions is an important aspect of the job market for those who do seek academic careers. The rate of tenure-program employment for those four to six years since receipt of Ph.D. fell from 26.6 percent in 1993 to 22.9 percent in 1997, which reflects both job opportunities in academia and alternate opportunities for employment. For example, one of the largest declines in tenure-program employment occurred in computer sciences (from 51.5 percent in 1993 to 33.3 percent in 1997) where other measures of labor market distress are low and organizations of computer science departments report difficulties recruiting faculty. The attractiveness of other employment may also be an explanation for drops in tenure-program rates in several engineering disciplines. The attractiveness of alternate employment, however, is a less likely explanation for the smaller drops in tenure-program employment rates in fields with other measures of distress, such as physics and mathematical sciences (both of which have large IOF rates) and the biological sciences (which have low unemployment and IOF rates, but have other indications of labor market distress).
By a strict definition of occupational titles, 15.0 percent of employed recent Ph.D. recipients were in occupations outside S&E, 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. recipients in non-S&E occupations said that their jobs were unrelated to their degrees. (See text table 3-9.) By field, these respondents ranged from 1.4 percent of recent psychology and computer science Ph.D. graduates to 6.3 percent of recent Ph.D. graduates in mathematical sciences.
Across all fields of degree, the median salary for recent S&E Ph.D. recipients was $41,000, a increase of 2.5 percent from 1995. By field, this ranges from a low of $32,000 in the biological sciences to a high of $68,000 in electrical engineering. Text table 3-10 shows the distribution of salaries for recent Ph.D. recipients by field of degree. For all Ph.D. recipients, those at the top 10 percent of the salary distribution (the 90th percentile) were paid $71,000. (See text table 3-10.) The 90th percentile salary varied by field from a low of $55,000 for sociology-anthropology to a high of $86,000 for computer science Ph.D. recipients. At the 10th percentile, representing the lowest-paid 10 percent among each field, salaries ranged from $16,000 for sociology-anthropology Ph.D. recipients to $45,000 for industrial engineering.
Salaries for recent S&E Ph.D. recipients by sector of employment are given in text table 3-11. The median salary for a postdoc one to three years after receipt of degree was $28,000about half the median for a recent Ph.D. recipient working for a private company ($60,000). Many of the salary differentials between different S&E fields are narrower when examined within sector of employment. For those in tenure-track positions, median salaries ranged from about $37,000 in mathematical sciences to $50,000 for computer S&E. At private for-profit companies, median salaries ranged from $43,000 for psychology to $72,000 for computer science.
Changes in median salaries for recent (defined here as one to five years after receipt of degree) bachelorís, masterís, and Ph.D. graduates are shown in text table 3-12. Across all S&E fields, median salaries for Ph.D. recipients rose by just 2.3 percent from 1995 to 1997compared with 11.1 percent for bachelorís and 10.0 percent for masterís degree graduates. To a considerable extent however, the median salary across all fields of Ph.D. was held down by relatively more rapid growth in Ph.D. production in lower-paying fields, such as the biological and social sciences. Much larger increases were found in most individual disciplines, including double-digit increases in physics (17.5 percent), electrical engineering (15.8 percent), computer sciences (11.7 percent), and economics (10.0 percent). Declines in median salaries were observed in civil engineering (Ė3.8 percent) and sociology-anthropology (Ė2.7 percent).
One indicator of the quality of employment available to recent graduates is simply their answer to this question: If you had the chance to do it over again, how likely is it that you would choose the same field of study for your highest degree? When this was asked of those with S&E degrees received 1Ė5 years after their previous degree, 16.6 percent of Ph.D. recipients said they were "not at all likely" as compared with 20.2 percent of those with S&E bachelorís degrees. (See text table 3-13.) This regret of field choice is lowest for recent Ph.D. recipients in computer sciences (6.8 percent) and electrical engineering (9.8 percent), and in the social sciences (12.5 percent). It is greatest in physics (24.4 percent), chemistry (23.9 percent), and mathematical sciences (22.4 percent).
A postdoctoral appointment (or postdoc) 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 has been used in the Survey of Doctorate Recipients to ask respondents about current and past postdoctorate positions they have held. Data and analyses on postdoctorates are often examined in relation 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. The increasing use of postdocs has been a focal point of discussions about many inter-related topicsthe early career paths for new Ph.D. scientists, the vocational adequacy of Ph.D. programs, and the labor market expectations of new Ph.D. recipients.
Science & Engineering Indicators Ė 1998 included an analysis of a one-time postdoctorate module in the 1995 Survey of Doctorate Recipients that showed a slow increase the use of postdocs in many disciplines over time. In addition, in physics and the biological sciences, the fields with the heaviest use of postdocs, median time in postdoc positions extended well beyond the one to two years found in most other fields.
Data from 1997 show a small decline in the percentage of all recent S&E Ph.D. recipients entering postdoc positionsfrom 32.7 percent of 1994 graduates in 1995 to 30.7 percent of 1996 graduates in 1997. At the same time, however, Ph.D. recipients in earlier graduation cohorts in these two fields show a similar propensity to be in postdocs as those with the same years since degree in 1995. Speculatively, something like this might be observed if new graduates were the most affected by improvements in labor market conditions. In fields other than physics or biological sciences, the postdoc rate one year after degree fell only slightly, from 21.2 percent in 1995 to 19.9 percent in 1997.
Postdocs in 1997 were asked to provide reasons they were in their current postdoctoral appointmentthe distribution of "primary reasons" given is shown in text table 3-14. Across all fields of degree, 17.2 percent gave "other employment not available" as the primary reason they were in a postdoc. Most respondents gave as primary reasons that a postdoc was generally expected for a career in their field (23.7 percent), that they were seeking additional training either in or outside of their Ph.D. field (20.0 and 13.3 percent), or other reasons more consistent with the stated training and apprenticeship functions of postdocs.
Of those in postdoctorate positions in April 1995, 38.0 percent were still in a postdoctorate position in April 1997. (See text table 3-15.) This is a small reduction from the 41.6 percent of 1993 postdocs that were still postdocs in 1995. (See Science and Engineering Indicators 1998.) Only 16.5 percent had moved from a postdoctorate to a tenure-track position at a four-year educational institution (up from 12.1 percent in 1995); 18.3 percent found other employment at an educational institution; 18.0 percent were at a for-profit firm; 7.7 percent were employed at a nonprofit institution or government; and 1.5 percent were unemployed.
No information is available on the career intentions of those in postdoc positions, but it is often assumed that a postdoc is most valued by academic departments at research universities. More postdocs, however, in each field transition to for-profit employment than obtain tenure-track positionsand many tenure-track positions are at schools where a research record obtained through a postdoc appointment may not be of central importance.
The size of the S&E labor force, its productivity, and opportunities for new S&E workers are all greatly affected by the age distribution and retirement patterns of the S&E labor force. For many decades, rapid increases in new entries to the S&E labor force led to a relatively young S&E labor force with only a small percentage near traditional retirement ages. This general picture is rapidly changing as the large number of individuals who earned S&E degrees in the late 1960s and early 1970s are moving into what is likely to be the latter part of their careers.
The possible effects of age distribution on scientific productivity are controversial. Increasing average ages may mean increased levels of experience and productivity among scientific workers. Others have argued that it can reduce the opportunities for younger scientists to perform independent work. Indeed, in many fields there is scientific folklore and some actual evidence indicate that the most creative research comes from the young. The ongoing research in cognitive aspects of aging and the sociology of science is relevant to this debate, but will not be reviewed here.
Age distributions for scientists and engineers in the labor force are affected by many factorsnet immigration, morbidity, and mortalitybut most of all by historical S&E degree production patterns. Age distributions for individuals with S&E degrees in 1997 are given by degree level and field of degree in appendix table 3-19. With the exception of new fields, such as computer sciences (where 70.0 percent of degree holders are under age 40), the greatest population density of individuals with S&E degrees occurs between ages 40 and 49. This can be seen in figure 3-12, which shows the age distribution of the S&E educated labor force broken down by level of degree. For all S&E degrees there is also a bump up in the age distribution at ages 25Ė29 representing 14.2 percent of S&E degree holders in the labor force. This bump up, however, appears to be largely caused by increased degree production in the social sciences (where 25- to 29-year-olds represent 17.7 percent of the total). In general, most of the S&E degreed labor force is their late 30s through early 50s.
This general pattern holds true even for those with doctorates in S&E. Although Ph.D. holders are somewhat older than other S&E degree holders, this circumstance is because of fewer Ph.D. holders in the younger age categories, given the time needed to obtain this degree. The greatest population density of S&E Ph.D. holders in the labor force occurs for 45- to 54-year-olds.
S&E Ph.D. holders employed in tenured or tenure-track positions in four-year institutions of higher education (26.9 percent of all S&E Ph.D. holders) are somewhat older than all S&E Ph.D. holders31.5 percent older than age 54 compared to 25.8 percent. (See figure 3-13.) The greatest population density of Ph.D. holders in these tenure programs occurs between ages 40 and 59. It is worth noting the sharp differences between the 55Ė59 and 60Ė64 age categories for both academic Ph.D. holders and the S&E Ph.D. population as a wholea 48 percent drop that is much steeper than for the bachelorís or masterís degreed S&E population.
At all degree levels and fields, only a small proportion of the S&E degreed labor force was near traditional retirement agesonly 13.6 percent overall were over age 54. This has several likely important and often overlooked effects on the future S&E labor force:
Retirement behavior can differ in complex ways from one individual to the next. Some individuals "retire" from a job while continuing to work full- or part-time, sometimes for the same employer. Others leave the labor force without a "retired" designation from some formal pension plan. Three different ways of thinking about changes in labor force involvement are summarized in text table 3-16 for S&E degree holdersleaving full-time employment, leaving the labor force, and retiring from a particular job.
By age 63, 50 percent of S&E bachelorís and masterís degree holders were not working full-time. For S&E Ph.D. holders, this 50 percent mark is not reached until three years later, at age 66. Longevity also differs by degree level with other measures. Half of S&E degree holders have left the labor force entirely by age 64 for bachelorís degree holders, by age 65 for masterís degree holders, and not until age 68 for Ph.D. holders. Formal retirement also occurs at somewhat higher ages for Ph.D. holdersmore than 50 percent of S&E bachelorís and masterís degree holderís have "retired" from some job by age 63, compared with age 65 for S&E Ph.D. holders.
Although many S&E degree holders who formally "retire" from one job continue to work full-time or part-time, this occurs most often among those under age 63. (See text table 3-17.) The drop in labor force participation among the "retired" is more pronounced for part-time workolder retired S&E workers are actually more likely to be working full-time than part-time. Retired Ph.D. scientists and engineers follow the same pattern, albeit with somewhat greater rates of post-retirement employment than bachelorís and masterís degree holders.
Movement out of full-time employment by S&E degree holders aged 55Ė70 is shown in figure 3-14. At all degree levels, the proportion of S&E degree holders who work full-time declines fairly steadily by age. After age 55, full-time employment by S&E doctorates becomes significantly greater than for bachelorís and masterís degree holders. At age 70, over 20 percent of S&E Ph.D. holders are working full-time, compared with 10 percent of bachelorís and masterís.
Academic employment may be one reason for somewhat slower retirement among Ph.D. holders. Text table 3-18 looks at the rate at which S&E Ph.D. holders leave full-time employment between 1995 and 1997 by sector of employment. Within each age group, a smaller proportion of S&E Ph.D. holders employed in 1995 at four-year colleges and universities, or by the government, left full-time employment than S&E Ph.D. holders as a whole, or those employed by for-profit companies.
While slower retirement for S&E Ph.D. holders, particularly those in academia, is significant and of some policy interest, it is important to recognize that this does not mean that academic or other Ph.D. holders seldom retire. Indeed, figure 3-14 shows that their retirement patterns are much more like those of bachelorís and masterís degree holders than they are differentretirement is just delayed two or three years. Even the two-year transition rates for academia in text table 3-18 shows more than a third of those aged 66Ė70 leaving full-time employment over a two-year period.
One reason academic Ph.D. retirement rates have been of interest has been a concern that the academic tenure system, combined with the end of mandatory retirement under U.S. antidiscrimination laws, could lead to continued employment of many less productive professors. Text table 3-19 compares two-year transition rates of leaving full-time employment for S&E Ph.D. holders employed full-time in 1995 at four-year institutions, by the number of articles they said they published within the previous five years. Within each age group, those writing six or more articles had a much lower transition rate out of full-time employment than those reporting fewer articles written. For those between the ages of 51 and 65, the transition rate for academics with zero articles was more than double the rate for those with six or more.