Page 139

C
EMPLOYMENT TRENDS AMONG SCIENTISTS AND ENGINEERS WITH GRADUATE DEGREES

Michael McGeary
Study Director, Committee on Science,
Engineering, and Public Policy

Contents

Page 140

Page 141

Page 142

OVERVIEW

This appendix has two parts: an original "cohort" analysis of data on the employment of recent science and engineering (S&E) PhDs since 1973 and a review of data on the employment of new science and engineering master's-degree recipients.

EMPLOYMENT TRENDS AMONG RECENT COHORTS OF SCIENCE AND ENGINEERING DOCTORATES

To understand better the career prospects of recipients of advanced degrees in science and engineering, a thorough knowledge of trends in the recent employment-related histories of new graduates is helpful. A rich database for such an analysis exists, and a preliminary study is reported in this appendix. The database is the Survey of Doctorate Recipients (see Box C-1 on page 148 for a description of the SDR database and its potential uses and limitations).

NSF publishes tables of data from the SDR on the entire population of U.S. scientists and engineers (the most recent was of the 1991 survey; see NSF, 1994d). The tables are examined below, but they do not provide information about recent science and engineering PhDs. For this study, OSEP was asked to produce two series of data tables: (1) tables on the employment activities of scientists and engineers who had received science and engineering PhDs in the 1-4 years before each survey, and (2) tables on the employment activities of those receiving science and engineering PhDs 5-8 years before each survey. This type of cohort analysis was apparently last done in response to the "new depression" in academic employment of the middle 1970s (see, e.g., NRC, 1983).

In this appendix, 4-year "classes" or cohorts are used to ensure minimal sample size when looking at specific fields, such as mathematics, chemistry, and biology. For example, the first set of tables gives information on those getting PhDs in 1969-1972 at the time of the 1973 survey, those getting PhDs in 1971-1974 at the time of the 1975 survey, and so on, through those who earned PhDs in 1987-1990 at the time of the 1991 survey.

The tables present data on workforce status (full-time, part-time, or unemployed), employment sector (academe, other education, business and industry, government, etc.), and tenure status of those employed in academe (4-year colleges, universities, and medical schools). There are tables and/or figures for the following:

Page 143

All scientists and engineers (excluding psychologists)

Physical scientists

Mathematicians
Computer scientists
Physicists/astronomers
Chemists
Earth/atmospheric/ocean scientists

Life scientists

Agricultural scientists
Medical scientists
Biological scientists

Engineers
Social scientists
Psychologists

Employment Status: Full-Time, Part-Time, and Not Employed

1-4 Years Out For those 1-4 years after receipt of PhD at the time of each survey since 1973, the SDR data (not reproduced here) show that unemployment rates and part-time employment rates have been low in all fields. Overall, the percentage of unemployed (defined as not working and looking for work) has varied between 1.2% and 1.6%. It was 1.5% in 1991 among those getting PhDs in 1987-1990. This trend, or lack of a trend, holds for each field. It does not rule out, however, substantial increases in unemployment since 1991.

The percentage not employed for all reasons (e.g., not looking, retired, or unemployed) has also been low and steady since 1973, about 3%.

The proportion of new PhDs employed part-time has increased by about 50%—from 2% in 1973 to 3.1% in 1991 (totaling about 1,100 of 58,000 in 1973 and 2,000 of 63,000 in 1991). This trend was shown in all science fields, except that part-time employment tripled among new social scientists, from 2% to 6%. Part-time employment did not increase among new engineering PhDs (it has been about 1% since 1973).

The category with the most significant change has been the proportion of recent doctorates employed in postdoctoral appointments. Just 8% of the 1969-1972 PhDs were postdoctorates in 1973, a percentage that increased steadily to 19% of the 1985-1988 PhDs in 1989 (the 1991 survey was delayed six months, so the postdoctorate percentage is lower—14%). Most postdoctorates were among new biology PhDs (2,000 in 1973 and 6,300 in 1989), and most of the rest were in physics/astronomy and chemistry (1,900 in 1973 and 2,800 in 1989).

5-8 Years Out For those 5-8 years out at the time of each survey, the pattern is similar except that few are in postdoctoral appointments. The percentage of unemployed is generally lower than for those 1-4 years out—between 0.7% and 1.3%—but it increased to 1.9% in 1991.

Page 144

The percentage not employed for other reasons was up a little in 1991, but 97% were still employed. The proportion employed part-time doubled to 3.2% in 1991 but was still low.

Conclusion If there are major employment changes or problems, they have arisen mostly since 1991. The employment statistics are all relatively good, but some were slightly worse in 1991 than in 1987 or 1989. Even that is difficult to interpret, because the survey procedures were changed in 1991 and the response rate was much higher.1

Employment Sector: Academe, Industry, Government

Figure C-1 shows where all science and engineering PhDs were employed 1-4 years after receiving their doctorates, and Figure C-2 shows where they were 4 years later (5-8 years after receiving their doctorates.); Tables C-1 and C-2 give the numbers and percentages on which the figures are based. Figure C-1 confirms the steadily growing proportion going into postdoctoral positions during the 1970s and 1980s (8% in 1973 and 19% in 1989, as noted above)2as well as the growing proportion going into business and industry (22% in 1973 and 29% in 1991). A smaller proportion went directly into academic employment (49% in 1973 and 40% in 1991)(presumably, some were going into postdoctoral positions instead).

Figure C-2 shows where science and engineering are 5-8 years after receiving their doctorates. Not surprisingly, few are in postdoctoral appointments, although the percentage increased from 2% to 3% over the period (but it increased from 4.8% in 1973 to 9.2% in 1989 among biologists 5-8 years out). The proportion employed in academe fell. More than half the 1969-1972 PhDs were employed in academe 5-8 years later, compared with 45% of the 19831986 graduates in 1991. This trend is the basis of a major conclusion of the report—that most new PhDs are employed outside academe.

Another way of looking at the trend is to compare the same "class" 1-4 years and 5-8 years out (Figures C-3 to C-5). Over the period, there is an increasing falloff from the number of those on postdoctorates and employed in academe in the first 1-4 years to the number employed in academe plus postdoctorates after 5-8 years (this is true, although to different degrees, across fields).

The growth sector was business and industry. 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 (Figures C-3 through C-5).

1 As the report was going to press, after this was written, NSF released the preliminary results of the 1993 SDR. The unemployment rates were up slightly but still low (see text Figure 2-5).

2 It is important to remember that some members of each 4-year cohort have already completed a postdoctoral assignment by the time of the survey. These figures are most useful for seeing trends.

Page 145

The employment patterns differ from field to field (see Figures C-6 through C-18 for those 1-4 years out and Figures C-19 through C-31 for those 5-8 years out). Some fields such as chemistry and engineering, have long had a high percentage of PhDs working in industry; others have had high percentages working in academe. Within the sciences, however, the trends have been similar—a smaller proportion going into academe and the federal government, and a greater proportion going into business and industry (there was also noticeable growth in nonprofit employment among social scientists). Engineering is one field in which the percentage working in academe after 5-8 years, although relatively low, has increased (from 29% in 1979 to 32% most recently), as has the percentage working in business and industry (from 50% to 58%). The federal government was the big loser, going from 13% (1,600 of the class of 19691972 in 1977) to 5% (about 500 of the class of 1983-1986 in 1991).

Tenure Status

Table C-2 has some interesting data on trends in tenure and tenure status within academe among those 5-8 years after receiving the PhD. The number of academics with tenure fell both absolutely and relatively over the period. In 1979, for example, 20% of all those who had received their PhDs 5-8 years before (1971-1974) had tenure, and another 14% had tenure-track positions. The percentage with tenure after 5-8 years fell steadily to 12% in 1989, while the percentage in tenure-track positions grew to 17%. In absolute terms, the number with tenure after 5-8 years fell from 12,000 in 1979 to 6,500 in 1989, while the number in tenure-track positions went from 8,000 to 9,000. (The 1991 survey increased the percentage with tenure to at least 14%; the extent to which the change in survey methods contributed to this change needs to be explored.)

In another way of looking at the data, the total number in the tenure system of those 5-8 years after receiving the PhD fell from 20,000 in the 1979 survey to about 16,000-17,000 in the 1989-1991 survey, and the proportion with tenure decreased relative to those still in tenure-track jobs.

Meanwhile, a fairly steady number and percentage of PhDs 5-8 years out had non-tenure system positions over the period (about 7,000, or 12% of all PhDs 5-8 years out).

The figures differ from field to field, although the general trends hold. Biology is interesting because the percentage with tenure, low in 1979 at 18% fell to 7% in 1989, while those in tenure-track positions or not in the tenure system stayed about the same. Presumably, that reflects the high and growing proportion of postdoctorates and the lengthening time before biology PhDs enter faculty positions, and begin their quest for tenure. In fact, as noted above, more than 9% of biologists were still in postdoctoral positions 5-8 years after receiving their PhDs. That in turn might be part of the explanation for the low and falling percentage of principal investigators under 35 years of age applying for National Institutes of Health research project grants, as reported in the recent National Research Council study of funding of young investigators (NRC, 1994a).

Page 146

Impact of Employment Choices of Recent PhDs

The new data presented here on science and engineering 1-4 and 5-8 years after receiving their doctorates give a sense of the career-related flow of new PhDs. The flows also have an impact on the overall stock of PhDs. More than 14,000 new doctorates were added each year to the total stock of PhDs working in the United States, and small shifts in employment patterns have a cumulative impact over time on the whole. Table C-3 presents data on the employment of all US PhDs at the time of each survey.

· With 14,000 a year being added to a stock that was relatively young in 1973, the total number of employed science and engineering PhDs increased from 220,000 in 1973 to 437,000 in 1991.

· Although the percentage in academe (four-year colleges, universities and medical schools) fell from 57% to 49%, the absolute number increased from 125,000 in 1973 to 195,000 in 1991 (an increase of 56%).

· The overall percentage of those in academe with tenure grew from 58% in 1975 to 62% in 1983 and fell to 55% in 1991.

· The percentage employed in business and industry increased from 24% to 36%, or from 53,000 to 157,000 (an increase of 196 percent).

Conclusions

1. More than 14,000 new PhDs in science and engineering have found employment each year since the early 1970s. Unemployment rates of those out for 1-4 years or 5-8 years have been low compared with other occupations and fairly stable, at least through 1991. That does not reflect events since 1991, and it does not mean that all found work in their fields or found jobs that they expected or wanted.

As a result of the steady output of doctoral science and engineering, the overall number of people with science and engineering PhDs from U.S. universities working in the United States has nearly doubled since 1973. Academe has absorbed a large number, 70,000, or about 3,900 a year. Business and industry have absorbed another 104,000, or 5,800 a year. The latter has been the growth sector for PhD employment, and is likely to remain so for future PhDs.

2. Year by year, the proportion of new PhDs going into academe for employment has fallen steadily, with business and industry increasing their share of PhD employment. It is important to note that academe, while losing share, is still an important employer of new PhDs, and any changes in the academic employment market would have a substantial, if slowly declining, impact on the career prospects of new PhDs.

Page 147

3. The increasing proportion of new PhDs taking postdoctoral appointments has been an important feature of some fields, especially biology (accounting for more than half) and some of the physical sciences. The percentage of those 5-8 years out, in postdoctoral positions, especially in biology, is increasing, indicating perhaps a lengthening of such positions or an increase in the practice of taking multiple postdoctorates. Those trends might account in part for the decreasing percentage of PhDs with tenure 5-8 years out, because it postpones beginning of an independent faculty career. Also, increasing numbers of postdoctorates are going into nonacademic employment.

4. The data do not show a substantial increase in part-time employment or in nontenure positions in academe, at least through 1991. However, a large percentage (12%) and number (7,000) of recent PhDs, all those 5-8 years out, have been in such positions through the period. Who are they and what are they doing?

5. This analysis should be extended by looking at data on such matters as salary (academe versus business versus government), primary activity (e.g., basic or applied research, development, R&D management, teaching, or consulting), and field-switching.

Page 148

Box C-1: Methodological note

Page 149

Page 150

NOTES: This table compares cohorts of scientists and engineers 1-4 years after receiving a doctorate from a US university (e.g., 1969-1972 PhDs in 1973, 1971-1974 PhDs in 1975, and so on). All fields are included except psychology (for field-level data, see Figures C-6 through C-18). All percentages are of the total population, including those not in the workforce.

Due to changes in survey procedures and timing in the 1991 survey, the 1991 estimates are not entirely comparable to those for the earlier survey years (see methodological note in Box C-1 for further information on comparability of 1991 estimates).

SOURCE: Special runs of data on employment status and employment sector of US doctoral scientists and engineers from the Survey of Doctorate Recipients.

Page 151

Page 152

NOTES: This table compares cohorts of scientists and engineers 5-8 years after receiving a doctorate from a US university (e.g., 1969-1972 PhDs in 1977, 1971-1974 PhDs in 1979, and so on). All fields are included except psychology (for field-level data, see Figures C-19 through C-31). All percentages are of the total population, including those not in the workforce.

Due to changes in survey procedures and timing in the 1991 survey, the 1991 estimates are not entirely comparable to those for the earlier survey years (see methodological note in Box C-1 for further information on comparability of 1991 estimates).

SOURCE: Special runs of data from the Survey of Doctorate Recipients on employment status and employment sector of US doctoral scientists and engineers.

Page 153

NOTES: All numbers in the table are estimates derived from a sample.

All doctoral scientists and engineers employed in a science or engineering (S&E) field were categorized by their field of employment when that information was available. When it was not, or when the employment field was other than science or engineering, doctorate holders were categorized by their field of degree.

SOURCE: Calculated from NSF, 1991, and NSF, 1994d.

Page 154

1/ This category was first introduced in 1987 to conform to other data series produced by NSF.

NOTE: All numbers in the table are estimates derived from a sample.

SOURCE: Calculated from NSF, 1991, and NSF, 1994d.

Page 155

FIGURE C-1 Employment of all scientists
and engineers 1-4 years after US PhD.

FIGURE C-2 Employment of all scientists
and engineers 5-8 years after US PhD.

SOURCE: Tables C-1 and C-2

NOTES: See notes to Table C-1 and methodological note in Box C-1 for important information on the comparability of 1991 estimates with earlier estimates.

Not employed includes those seeking work (i.e., unemployed), not seeking work, retired, otherwise not working, or not reporting employment status.

Other employment includes other education (junior colleges, 2-year colleges, technical institutes, and elementary, middle, and secondary schools); state and local governments; hospitals and clinics; private foundations and other nonprofit organizations; other employers; and those who did not respond to the employment-sector question.

Business/industry includes self-employed.

Academe includes those employed at 4-year colleges, universities, and medical schools (including university-affiliated hospitals and medical centers).

Page 156

FIGURE C-3 1969-1972 science and engineering
PhD recipients 1-4 and 5-8 years later.

FIGURE C-4 1977-1980 science and engineering
PhD recipients 1-4 and 5-8 years later.

FIGURE C-5 1983-1986 science and engineering
PhD recipients 1-4 and 5-8 years later.

SOURCE: Tables C-1 and C-2.

NOTES: See notes to Table C-1.

Page 157

FIGURES C-6 THROUGH C-31

Page 158

FIGURE C-6 Employment of physical
scientists 1-4 years after US PhD.

FIGURE C-7 Employment of mathematicians
1-4 years after US PhD.

FIGURE C-8 Employment of computer
scientists 1-4 years after US PhD.

FIGURE C-9 Employment of
physicists/astronomers 1-4 years after US PhD.

FIGURE C-10 Employment of chemists 1-4
years after US PhD.

FIGURE C-11 Employment of environmental
scientists 1-4 years after US PhD.

Page 159

FIGURE C-12 Employment of life scientists
1-4 years after US PhD.

FIGURE C-13 Employment of agricultural
scientists 1-4 years after US PhD.

FIGURE C-14 Employment of medical
scientists 1-4 years after US PhD.

FIGURE C-15 Employment of biological
scientists 1-4 years after US PhD.

Page 160

FIGURE C-16 Employment of social scientists 1-4 years after US PhD.

FIGURE C-17 Employment of psychologists 1-4 years after US PhD.

Page 161

FIGURE C-18 Employment of engineers 1-4 years after US PhD.

Page 162

FIGURE C-19 Employment of physical
scientists 5-8 years after US PhD.

FIGURE C-20 Employment of mathematicians
5-8 years after US PhD.

FIGURE C-21 Employment of computer
scientists 5-8 years after US PhD.

FIGURE C-22 Employment of
physicists/astronomers 5-8 years after US PhD.

FIGURE C-23 Employment of chemists 5-8
years after US PhD.

FIGURE C-24 Employment of environmental
scientists 5-8 years after US PhD.

Page 163

FIGURE C-25 Employment of life scientists
5-8 years after US PhD.

FIGURE C-26 Employment of agricultural
scientists 5-8 years after US PhD.

FIGURE C-27 Employment of medical
scientists 5-8 years after US PhD.

FIGURE C-28 Employment of biological
scientists 5-8 years after US PhD.

Page 164

FIGURE C-29 Employment of social scientists 5-8 years after US PhD.

FIGURE C-30 Employment of psychologists 5-8 years after US PhD.

Page 165

FIGURE C-31 Employment of engineers 5-8 years after US PhD.

Page 166

EMPLOYMENT OF RECENT SCIENCE AND ENGINEERING MASTER'S DEGREE RECIPIENTS

The National Science Foundation conducts biennial surveys of recent recipients of master's and bachelor's degrees in various science or engineering fields. The latest published data are from the survey administered in the spring of 1990 to a sample who received science and engineering degrees during the 1987-1988 (1988) and 1988-1989 (1989) academic years (NSF, 1992a). Surveys conducted in 1982 (NSF, 1984), 1984 (NSF, 1986), 1986 (NSF, 1987), and 1988 (NSF, 1990a) are also reported here. A more recent survey was fielded in 1993 and is scheduled to be published shortly.

The NSF survey of recent college graduates and master's-degree recipients collects information on demographic and education characteristics and on early career-development experiences, such as employment status, reasons for unemployment, and attributes of employment including occupational classification, major activity, and salary. The survey is based on a nationally representative sample and is used to derive national estimates of the numbers and demographic, education, and employment characteristics of recent graduates in science and engineering.

How Many Science and Engineering Master's Degree Recipients Continue Graduate Study?

The number of new science and engineering master's-degree recipients in 1989 continuing as full-time students in 1990 was 16,200, or about 23% of the estimated 70,400 who received master's degrees in 1989 (another 4,900, or 7%, were part-time students)(Table C-4). The comparable percentage and number of continuing students among new master's-degree recipients in the 1980s surveys were about 21% and 10,000 - 13,000.

Page 167

SOURCE: Calculated from Table B-28 in NSF, 1984; Table B-28 in NSF, 1986; Table B-31 in NSF, 1987; Table B-31 in NSF, 1990a; and Table B-31 in NSF, 1992b.

Continued Education of New Science and Engineering Master's-Degree Recipients

The tendency to continue graduate studies full-time varies across fields but has increased in most fields since 1982. Barely 7% of those receiving master's degrees in computer science in 1989 were full-time students the next year, compared with 60% of those with new physics/astronomy master's degrees (Table C-5). The proportion of new master's degree recipients going to graduate school part-time also varies from field to field but has gone down since 1982. The ratio of full-time to part-time graduate students a year after receipt of master's degrees also varies by field. In fields where the master's degree is usually considered the working degree, the ratio is more nearly equal (e.g., 2.2:1 in engineering and 1:1 in computer science and technology) than in fields where the PhD is considered the minimal professional credential (e.g., 3.6:1 in physics/astronomy, 22.6:1 in biology, and apparently higher in chemistry).

Page 168

a Too few cases to report.

SOURCES: Calculated from Table B-28 in NSF, 1984, for 1981 data; and Table B-31 in NSF, 1992b, for 1989 data.

Page 169

How Many Recent Science and Engineering Master's-Degree Recipients Are Unemployed?

Unemployment rates among recent science and engineering master's-degree recipients in the labor force are low. In 1990, among those awarded master's degrees in 1988 and 1998, unemployment was 1.8% overall, ranging from a low of 1.1% of mathematicians and statisticians to a high of 3.6% of psychologists (Table C-6). In most fields, the unemployment rate among recent master's-degree recipients was lower in 1990 than in 1982 and 1984 (in the aftermath of the 1980-1982 recession), but slightly higher than in 1988.

a Full-time graduate students are excluded.

b No rate computed for groups in which labor force is smaller than 1,500.

SOURCES: Calculated from Table B-45 in NSF, 1984; Table B-45 in NSF 1986; Table B-49 in NSF, 1987; Table B-47 in NSF, 1990a; and Table B-47 in NSF, 1992b.

Page 170

How Many New Science and Engineering Master's-Degree Recipients Are Employed in Their Fields or Other Science and Engineering Occupations?

Although unemployment rates are comparatively low for those with master's degrees, not all are working in the fields of their degrees or in other science and engineering occupations. Overall, in 1990, fewer than two-thirds of science and engineering master's degree recipients were working in the same fields as their degrees, and fewer than one-fifth were working in other science and engineering occupations (Table C-7). Exactly 17% were not employed in science and engineering occupations. This split varied by field, however, with 7.5% of engineers and 8% of computer scientists working outside science and engineering, compared with 37% of social scientists and 46% of psychologists. Only one-fourth of physicists and astronomers were working in their fields (25% were working in computer science and ''other physical sciences," and 25% in engineering occupations) (NSF, 1992b:Table B-39).

a Exclusive of full-time graduate students.

NOTE: Percentages may not add to 100 because of rounding.

SOURCE: Calculated from Table B-39 in NSF, 1992b.

Page 171

Who Employs New Science and Engineering Master's-Degree Recipients?

In the last decade, nearly 60% of employed new science and engineering master's degree recipients were working in the private sector (Table C-8). Less than 20% were employed by educational institutions, split about evenly between four-year colleges and universities and other institutions (two-year colleges and elementary and secondary schools).

a Full-time graduate students are excluded.

b Too few cases to report.

SOURCES: Calculated from Table B-32 in NSF, 1984; Table B-32 in NSF, 1986; Table B-36 in NSF, 1987; Table B-36 in NSF, 1990a; and Table B-36 in NSF 1992b.

Page 172

Field-to-Field Differences in Employment Sector

The tendency to work in industry or for other types of employers varies by field. In 1990, for example, more than three-quarters of computer scientists worked in the private sector, compared with less than a third of life scientists (Table C-9).

a Full-time graduate students are excluded.

b Too few cases to report.

SOURCE: Calculated from Table B-36 in NSF, 1992b.