Skip all navigation and go to page content

Chapter 3. Science and Engineering Labor Force

Demographics


Age and Retirement

The age distribution and retirement patterns of the S&E labor force affect its size, productivity, and the opportunities it offers for new S&E workers. For many decades, rapid increases in new entries into the workforce created a relatively young pool of workers, with only a small percentage near traditional retirement age. Now, individuals who earned S&E degrees in the late 1960s and early 1970s are moving into the later part of their careers.

The increasing average age of S&E workers may mean increased experience and greater productivity among them. However, it could also reduce opportunities for younger researchers to make productive contributions by working independently. In many scientific fields, folklore and empirical evidence indicate that the most creative research comes from younger people (Stephan and Levin 1992).

Aside from the possible effects on productivity, early career opportunities, and, perhaps, the culture within some scientific fields, the age structure of the S&E labor force has important implications for its growth rate. This section does not attempt to project future S&E labor market trends; however, it posits some general conclusions. Absent changes in degree production, retirement patterns, or immigration, the number of S&E-trained workers in the labor force will continue to grow for some time, but the growth rate may slow considerably as an increasing proportion of the S&E labor force reaches traditional retirement age. With slowing growth, the average age of the S&E labor force will increase.

Age Distribution of the S&E Workforce
Net immigration, morbidity, mortality, and, most of all, historical S&E degree production patterns affect the age distribution of scientists and engineers in the workforce. With the exception of new fields such as computer sciences (in which 56% of degree holders are younger than age 40), the greatest population density of individuals with S&E degrees occurs between the ages of 40 and 49. Figure 3-21 shows the age distribution of the labor force with S&E degrees broken out by level of degree. In general, the majority of individuals in the labor force with S&E degrees are in their late thirties through their early fifties, with the largest group at ages 40–44. More than half of workers with S&E degrees are age 40 or older, and the 40–44 age group is more than twice as large as the 60–64 age group.

This general pattern also holds for individuals with S&E doctorates. Because of the length of time needed to obtain a doctorate, those who hold these degrees are somewhat older than individuals who have less advanced S&E degrees. The greatest population density of S&E doctorate holders occurs between the ages of 40 and 54. This can be seen most easily in figure 3-22 , which compares the age distribution of S&E degree holders in the labor force at each level of degree, and in figure 3-23 , which shows the cumulative age distribution for individuals at each degree level. Even if one takes into account the somewhat older retirement ages of doctorate holders, a much larger proportion of S&E doctorate holders are near traditional retirement ages than are individuals with either S&E bachelor's or master's degrees.

Figure 3-24 , which compares the age distributions of S&E doctorate holders in 1993 and 2003, highlights the extent of the shift in the age structure of the S&E labor force. S&E doctorate holders under age 35 are about the same proportion of the S&E doctorate holders in the total labor force in both years. However, over the decade, the 35–54 age group became a much smaller proportion of the doctoral-level S&E labor force. What grew was the proportion of S&E doctorate holders age 55 and older.

Across all degree levels and fields, 26.4% of the labor force with S&E degrees is older than age 50. The proportion ranges from 15% of individuals with their highest degree in computer sciences to 41% of individuals with their highest degree in geosciences (figure 3-25 ).

Altogether, the age distribution of S&E-educated individuals suggests the following likely effects on the future of the S&E labor force:

  • Barring large changes in degree production, retirement rates, or immigration, the number of trained scientists and engineers in the labor force will continue to increase, because the number of individuals currently receiving S&E degrees exceeds the number of workers with S&E degrees nearing traditional retirement age.
  • However, unless large increases in degree production occur, the average age of workers with S&E degrees will rise.
  • Barring large reductions in retirement rates, the total number of retirements among workers with S&E degrees will increase over the next 20 years.

Taken together, these factors suggest a slower growing and older S&E labor force. Both trends would be accentuated if either new degree production were to drop or immigration were to slow, both concerns raised by a 2003 report of the Committee on Education and Human Resources Task Force on National Workforce Policies for Science and Engineering of the National Science Board (NSB 2003).

S&E Workforce Retirement Patterns
The retirement behavior of individuals can differ in complex ways. Some individuals retire from one job and continue to work part time or even full time at another position, sometimes even for the same employer. Others leave the workforce without a retired designation from a formal pension plan. Table 3-11 summarizes three ways of looking at changes in workforce involvement for S&E degree holders: leaving full-time employment, leaving the workforce, and retiring from a particular job.

By age 61, slightly more than 50% of those with an S&E bachelor's degree as their highest degree are no longer working full time. The age at which at least half of S&E degree holders no longer work full time increases by degree level—to age 62 at the master's level and age 66 at the doctoral level. Longevity also differs by degree level when measuring the number of individuals who leave the workforce entirely: half of all S&E bachelor's degree holders left the workforce entirely by age 65, compared with S&E master's degree and doctorate holders who left the workforce at ages 66 and 69, respectively. Although many S&E degree holders who formally retire from one job continue to work full time or part time, formal retirement occurs at similar ages for all levels of degree holders: more than 50% of bachelor's, master's, and doctoral degree holders have formally retired from jobs by age 65, 66, and 67, respectively.

Figure 3-26 shows data on S&E degree holders working full time at ages 55–69. For all degree levels, the proportion of S&E degree holders who work full time declines fairly steadily by age, but after age 55, full-time employment for doctorate holders becomes significantly greater than for bachelor's and master's degree holders. At age 69, 27% of doctorate holders work full time, compared with 16% of bachelor's degree recipients.

Table 3-12 shows the rates at which holders of U.S. S&E doctorates left full-time employment, by sector of employment, between October 2003 and April 2006. For every age group, the retirement rates for S&E doctorate holders were slightly higher for those working in the private sector than those employed in education or government. 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 individuals younger than age 63 (table 3-13 ). However, of retired S&E degree holders age 71 to 75, only 12% of bachelor's degree holders keep working either full time or part time, 17% of master's degree holders, and 19% of doctorate holders.


Women and Minorities in S&E

An important part of the growth of the S&E labor force comes from the increased presence of women and ethnic minorities. In 2006, white males constituted 58% of those in the labor force over age 50 whose highest degree was in S&E. Among those under age 30, only 35% were white males (NSF/SRS 2006). This represents both a change in the composition of the total U.S. labor force and a growth in the participation of women and minorities in S&E.

Both women and underrepresented ethnic minorities have shown steady growth in their proportion of the S&E labor force (see figures 3-27 and 3-28 , which look at sex and ethnic representation within S&E occupations).

Representation of Women in S&E
Women constituted more than one-fourth (26%) of the college-educated workforce in S&E occupations and two-fifths (40%) of those with S&E degrees in 2006, according to NSF's SESTAT data.

Census data on S&E occupations from 1980 to 2007 show the number of women in S&E occupations rising from 12% to 27% over those 27 years (figure 3-27). Figures 3-29 and 3-30 show the growth in the number of women with education in S&E for different graduation cohorts and broad fields of degree. The notable exception is in computer and mathematical sciences at the bachelor's degree level, where the proportion of women in the workforce is lower for 2002–05 graduates (27%) than it is for 1972–76 graduates (35%). In contrast, the proportion of women in the most recent bachelor's degree cohorts in both the social sciences and the life sciences has risen to above 60%. Among S&E doctorate holders in the workforce, the proportion of women is generally higher in more recent cohorts, including the computer and mathematical sciences.

Age Distribution and Experience. On average, women in the S&E workforce are younger than men (figures 3-31 and 3-32 ). Forty-six percent of women and 31% of men employed in science and engineering in 2003 received their degrees within the previous 10 years. The difference is even more profound at the doctoral level, which has a much greater concentration of women in their late thirties. Consequently, a much larger proportion of male scientists and engineers at all degree levels, but particularly at the doctorate level, will reach traditional retirement age during the next decade. This will affect sex ratios and potentially the number of female scientists in senior-level positions.

Unemployment. Unemployment rates in 2006 were somewhat higher for women in S&E occupations than for men: 2.2% of men and 2.9% of women were unemployed. In contrast, the unemployment rate in 1993 was 2.7% for men and 2.1% for women (table 3-14 ).

Representation of Racial and Ethnic Minorities in S&E
With the exception of Asians/Pacific Islanders, racial and ethnic minorities represent only a small proportion of those employed in S&E occupations in the United States. Collectively, blacks, Hispanics, and other ethnic groups (the latter category includes American Indians/Alaska Natives) constitute 24% of the total U.S. population, 13% of college graduates, and 10% of college-educated individuals employed in S&E occupations.

Conversely, Asians/Pacific Islanders, despite constituting only 5% of the U.S. population, accounted for 7% of college graduates and 14% of those employed in S&E occupations in 2003. Although most (82%) Asians/Pacific Islanders in S&E occupations were foreign born, those born in the United States were also more highly represented in S&E than in the total workforce.

Age Distribution. As is the case for women, underrepresented racial and ethnic minorities in the S&E workforce are much younger than non-Hispanic whites in the same S&E jobs (figure 3-33 ), and this difference is even more pronounced for doctorate holders in S&E occupations (figure 3-34 ). This finding could point to an upcoming shift in the overall composition of the S&E workforce. In the near future, a much greater proportion of non-Hispanic white doctorate holders in S&E occupations will be reaching traditional retirement ages. This circumstance could signal a more rapid increase in the number of non-Hispanic white doctorate holders who will retire or otherwise leave S&E employment. On the other hand, Asian/Pacific Islander doctorate holders in S&E occupations (measured by race and not by place of birth) are on average the youngest racial/ethnic group, and thus the least likely to have large numbers of retirees.

Salary Differentials for Women and Minorities
Trends in Median Salaries. Women and members of underrepresented minority groups have generally lower earnings than their male and nonminority counterparts. However, differences in average age, work experience, fields of degree, sector of employment, and other characteristics can make direct comparison of salary and earnings statistics misleading. This section discusses these income gaps and explores some of the underlying factors that may affect them.

Factors Influencing Salary Differentials. Regression analysis is a statistical method that can be used to examine salary and other differences simultaneously.[7] Although this type of analysis can provide insight, it cannot give definitive answers to questions about the openness of S&E to women and minorities. The most basic reason is that no labor force survey ever captures information on all characteristics that may affect compensation.

Figures 3-35 and 3-36 show estimates of salary differences for different groups after controlling for several individual characteristics. Differences in mean annual salary are substantial when comparing all individuals with S&E degrees by level of degree only.

In 2006, women with S&E bachelor's degrees working full time had mean salaries that were 36.2% less than those of their male counterparts. Likewise, full-time salaries of blacks, Hispanics, and individuals in other underrepresented ethnic groups with S&E bachelor's degrees were 25.8% less than those of non-Hispanic whites and Asians/Pacific Islanders with S&E bachelor's degrees.[8] While still substantial, these salary differentials decrease as level of degree increases for both women and ethnic minorities, reaching 21.1% and 15.0% respectively.

Effects of Age and Years Since Degree. On average, women and members of underrepresented minority groups are younger than their counterparts in most S&E fields. Controlling for differences in both age and years since receipt of degree reduces the estimated salary differential for both women and minorities at every degree level.

For women, it reduces salary differentials by about one-third at the bachelor's and master's degree levels, and by about half at the doctorate level.[9] Statistical controls may make less difference at lower degree levels because similar proportions of men and women with S&E degrees are in mid-career, but a larger proportion of men are at older ages, where salaries begin to decline.

For underrepresented ethnic minorities, controlling for age and years since degree produces proportionally larger reductions in salary differentials than is the case for women. Introducing these controls reduces salary differentials between underrepresented minorities and both non-Hispanic whites and Asians/Pacific Islanders by more than half at all degree levels.

Effects of Field of Degree on Salary Differentials. Controlling for field of degree in addition to age and years since degree reduces the estimated salary differentials for women with S&E degrees to –12.0% at the bachelor's degree level and to –7.6% at the doctorate level.[10] These reductions generally reflect the greater concentration of women in the lower-paying social and life sciences as opposed to engineering and computer sciences.

Field of degree is also associated with reduction of estimated salary differentials for underrepresented ethnic groups. Controlling for field of degree further reduces salary differentials to –9.1% for individuals with S&E bachelor's degrees and to –5.5% for individuals with S&E doctorates. At the doctoral level, field of degree, age, and years since degree together account for two-thirds of salary differentials for underrepresented ethnic groups.

Effects of Occupation and Employer Characteristics on Salary Differentials. Occupation and employer characteristics affect compensation.[11] Academic and nonprofit employers typically pay less for the same skills than employers pay in the private sector, and government compensation falls somewhere between the two groups. Other factors affecting salary are the sector of the economy, the U.S. region where a person works, and whether the person is working in S&E or in R&D. However, occupation and employer characteristics may not be determined solely by individual choice; they may also in part reflect an individual's career success.

When comparing women with men and underrepresented ethnic groups with non-Hispanic whites and Asians/Pacific Islanders, controlling for occupation and employer further reduces salary differentials. At the doctoral level, controlling for occupation leaves no statistically significant difference between the salaries of underrepresented ethnic groups compared with non-Hispanic whites and Asians/Pacific Islanders.

Effects of Family and Personal Characteristics on Salary Differentials. Marital status, the presence of children, parental education, and other personal characteristics are often associated with differences in compensation. Although these differences may involve discrimination, they may also reflect many subtle individual differences that can affect work productivity.[12] For example, having highly educated parents is associated with higher salaries for individuals of all ethnicities and both sexes. It may well be associated with greater academic achievement not directly measured in these data; alternatively, it may be associated with family and personal networks that are conducive to career success. In any event, for many individuals in many ethnic groups, historical discrimination probably affected parents' educational opportunities and achievement.

Controlling for these additional characteristics changes salary differentials only slightly for each group and degree level.[13] An additional issue for the wage differentials of women, however, is that family and child variables often have different effects for men and women. In these estimates, both marriage and children are associated with higher salaries for men with S&E degrees at all levels, but have a negligible association with women's earnings. Allowing for these differences in sex effects reduces the salary differential at the bachelor's degree level to 4.7% and leaves no statistically significant difference in salary at the master's degree and doctorate levels.

Notes

[7] Specifically presented here are coefficients from linear regressions using the 2003 SESTAT database of individual characteristics on the natural log of reported full-time annual salary as of October 2003.
[8] Underrepresented ethnic group, as used here, includes individuals who reported their race as black, American Indian/Alaska Native, of Hispanic origin, or other ethnicity.
[9] In the regression equation, this is the form: age1, age2, age3, age4; years since highest degree (YSD)1, YSD2, YSD3, YSD4.
[10] The regressions included 20 dummy variables for SESTAT field-of-degree categories (out of 21 S&E fields; the excluded category was "other social science").
[11] Variables added here include 34 SESTAT occupational groups (excluding "other non-S&E"), whether individuals worked in R&D, the employer's U.S. census region, and the sector of the economy.
[12] Variables added here include dummy variables for marriage, number of children in the household younger than 18, whether the father had a bachelor's degree, whether either parent had a graduate degree, citizenship, nativity, and age at receipt of first bachelor's degree minus 20. Sex and ethnic minority variables are included in all regression equations.
[13] This may be because differences between groups in many of these family and personal characteristics are not large. It is also possible that variations in these characteristics correlate with those in other controls already in the statistical model and in that sense have already been taken into account.
 

Science and Engineering Indicators 2010   Arlington, VA (NSB 10-01) | January 2010

Close