Skip all navigation and go to page content

Chapter 3. Science and Engineering Labor Force

Demographics of the S&E Workforce

This section describes the demographic composition of the U.S. S&E workforce by sex, race/ethnicity, foreign origin, and age. It also addresses the relationship between workforce demographics and selected indicators of labor force rewards and participation.

The section begins with a focus on differences by sex among workers in S&E occupations and among S&E degree holders. Similar comparisons will be made across race/ethnicity categories. Historically, in the United States, very high proportions of workers in S&E occupations have been male and white (non-Hispanic). Engineering and physical science occupations have had particularly low concentrations of women and of members of most underrepresented minority groups (i.e., blacks, Hispanics, American Indians, and Alaska Natives), both relative to the concentrations of these groups in other occupational areas and relative to their representation in the population in general. However, both women and minorities increasingly have been entering a wide range of S&E occupations. Asians have also been increasing their participation in S&E occupations, although with concentrations in areas different from women and underrepresented minorities. This section documents, across S&E occupations, the extent to which the numbers and the share of workers who are women, underrepresented minorities, and Asians have risen, and provides indicators of their contemporary levels of participation.

The presentation of indicators of levels of participation will be followed by an analysis of the relationship between wage differences and demographic factors. Historically, women and minorities in S&E occupations have received lower salaries than white men. This section will provide data on contemporary salary differences as well as findings regarding how various factors contribute to these differences.

This discussion of wage differences will be followed by a presentation of indicators pertaining to S&E immigration trends. Increasing global competition for S&E workers and changes in economic conditions influence levels of immigration. This section describes recent trends in immigration of S&E workers that can be compared with other factors (like economic growth). Indicators are collected from population data from the U.S. Census Bureau and visa data from the U.S. Citizenship and Immigration Service, as well as S&E workforce data from the NSF SESTAT data system. Data from the Survey of Earned Doctorates will be presented to capture stay rates: rates at which noncitizen recipients of U.S. S&E doctoral degrees remain in the United States.

The demographics section ends with a presentation of indicators of the aging of the S&E workforce as the baby-boom generation moves toward retirement age. The high concentration of workers over age 50 suggests that the S&E workforce will soon experience high levels of turnover. Thus, indicators will also be presented pertaining to levels of workforce participation and engagement of individuals at the ages near the end of their career cycle.

Sex Differences in the S&E Workforce

Sex Differences in S&E Occupations

Historically, men in S&E occupations have outnumbered women by wide margins. Yet the number of women in these occupations has been on the rise, increasing over the past two decades by more than half-a-million workers. These recent increases in the number of women have narrowed overall disparities by sex, but only modestly. In 2008, overall disparities remained pronounced, with women constituting 27% of workers, only a slightly higher share than in the previous decade when women made up 23% of workers (figure 3-28).

Sex disparities vary across occupations (appendix table 3-9). The most extreme disparities are within engineering, where women constituted 13% of the workforce in 2008. Among large engineering occupations, the disparity between men and women is greatest among mechanical engineers, with men outnumbering women by more than 12 to 1. Other large engineering occupations in which women account for as few as 9% of workers include electrical and electronics engineers and aerospace, aeronautical, and astronautical engineers.

Both computer and mathematical scientists (26% women) and physical scientists (30% women) are disproportionately male. Within physical science occupations, physicists have the largest imbalance by sex. Within computer and mathematical scientist occupations, the largest component, computer and information scientists (25% women), is the most imbalanced. Mathematical scientists (45% women) are much closer to parity.

Sex parity in participation was nearly achieved by 2008 among biological and medical scientists (51% women). With 53% of women in the field in 2008, parity in the social sciences has been long established. Occupations within social sciences, however, vary with respect to the proportion of female workers. For example, women account for slightly less than one-third of economists, but more than two-thirds of psychologists. Psychology, with about 185,000 total workers, is the only large S&E occupation with substantially more women than men.

The number of women working in each occupational area has risen since the early 1990s. Growth has been strongest in the biological and related sciences, where the number of female workers doubled between 1993 and 2008. This rate of growth has far outstripped that of men in these occupations, thus women's share of workers has also increased (from 34% in 1993 to 45% in 2008, see figure 3-28). During the same period, women have also increased their share among workers in engineering (from 9% to 13%) and in the physical sciences (from 21% to 30%). In these two occupational areas, women's increased share emerged as women's numbers in the workforce expanded (roughly by 60%) but men's numbers did not, remaining roughly similar between 1993 and 2008.

In social science occupations, the growth in women's participation has occurred at levels similar to those in engineering and the physical sciences. However, men's participation in these occupations has grown at similar levels and, therefore, the balance between men and women has changed little.

With 230,000 more female computer and mathematical scientists in 2008 than in 1993, women have added more workers in this area than in any of the other S&E occupations. The rate of growth of women in this area is also higher than in any other area, except for life scientists. However, unlike in the other four areas, men's rate of growth in this occupational area is higher than women's. Thus, women's share of this occupation has been declining. From 1993 to 2008 women's share of computer and mathematical scientists dropped from 31% to 26%, making the sex disparity here even greater than in physical science occupations. The declining share of women in the computer and mathematical science occupations reflects increasing disparities in participation among those whose highest degree is at the bachelor's degree level. Among those with a doctoral degree, women's share of workers in computer science occupations increased from 13% to 18% over this period.

Sex Differences in Age and Racial/Ethnic Groups

With the recent, greater growth among women than among men in S&E occupations, women in the field tend to be somewhat younger than the men (table 3-24). Age disparities are greatest among life scientists, physical scientists, and engineers, where women's participation levels have been increasing relative to men's. Age disparities are small among computer and mathematical scientists, where women have lost ground relative to men in levels of participation. Overall, in 2008 28% of men working in S&E occupations were over age 50 compared with 22% of women. Only 13% of men were younger than 30, but 17% of women were. The median age of women in S&E occupations was 41 years compared with 43 years among men.

Women in S&E occupations were more likely than men to be classified as an American Indian/Alaska Native, black, Hispanic, Native Hawaiian/Pacific Islander or of two or more races. In 2008, 14% of women in S&E occupations identified themselves within one of these groups compared with 10% of men (appendix table 3-10). Neither occupational area nor age explains the increased likelihood for women to be from a minority group, and less likely to be white. Women are more likely to be minorities within all five broad occupational areas whether or not age is controlled.

Sex Differences Among S&E Degree Holders

Sex disparities among the general U.S. workforce with S&E degrees are somewhat smaller than disparities within S&E occupations. In 2008, among individuals with their highest degree in an S&E field, women constituted 38% of those who were employed, up from 31% in 1993. Over the same period, the share of women among unemployed workers with an S&E degree rose more dramatically, from 34% to 45%. Among those out of the labor market, the share of women rose from 46% to 50%.

At every age, women with their highest degree in an S&E field are more likely than men to be out of the labor market (figure 3-29). However, at typical ages for career entry and exit (before age 30 and after age 60) these differences are modest. The sex disparity in the likelihood of being out of the labor market is particularly pronounced in the middle years of the career cycle. Between ages 30 and 55, 16.1% of women were out of the labor market compared with 2.2% of men.

Many women between ages 30 and 55 with S&E degrees who were not in the labor market identified family reasons as an important factor: 69% of women reported that family was a factor compared with 25% of men. Within this age range, women were also much more likely than men to report that they did not need to work or did not want to work (46% of women and 26% of men).

Sex Differences in Degree Fields and Degree Levels

With respect to the proportion of men and women among S&E highest degree holders, the pattern of variation among degree fields echoes the pattern of variation among occupations associated with those fields (see appendix table 3-11). In 2008, more than half (54%) of degree holders in the social science fields were women, as were nearly half (46%) of those with a degree in the biological and related sciences. Men outnumbered women among computer sciences and mathematics degree holders (31% women) and among physical science degree holders (28% women). Disparities, however, were greatest among those with a degree in engineering, where only 13% of degree holders were women. In all fields except computer and mathematical sciences, the share of women with degrees in the workforce has been increasing over the past two decades. In computer science and mathematics this share has remained flat.

Sex differences are not limited to the field of degree, but also to the level of the S&E degree. Men in the workforce are more likely to have a more advanced S&E degree. For example, women accounted for 38% of those whose highest degree in S&E is at the bachelor's level but 29% of workers whose highest degree in S&E is at the doctoral level (figure 3-30). At the doctoral degree level, however, women's share has been steadily increasing. Women's share of S&E bachelor's degree holders in the workforce has also been rising since the early 1990s, but in 2008 this share was not larger than it had been in 2006.

Working men and women with S&E degrees also differ in the extent to which they are employed in the same field as their S&E degree. However, this disparity is largely the result of women having a high concentration in the two degree areas—social sciences and life sciences—where degree holders most often work outside of S&E occupations.

In 2008, across all degree areas, 21% of women with a highest degree in S&E compared with 35% of men were employed in the field in which they earned their degree (appendix table 3-12). About 26% of women were working in an S&E occupation compared with 45% of men. However, within most degree areas, a similar proportion of men and women work in an occupation that matches their degree field, and similar proportions work in non-S&E jobs. Computer and mathematical science fields are exceptions, where men are more likely to work in an occupation that matches their degree field.

Racial/Ethnic Differences in the S&E Workforce

This section addresses the level of diversity in science and engineering by describing the cross-cutting social categories of race and Hispanic status. Like the preceding section, this section draws on data from the NSF science and engineering labor force surveys to report on levels of participation in science and engineering: first, across occupations, and next, across the overall workforce with science and engineering degrees.

Whether defined by occupation, S&E degree, or the combined criteria used in SESTAT, the majority of scientists and engineers in the United States are non-Hispanic white. The next largest group of scientists and engineers are Asians, who have been increasing their share in the S&E field since the early 1980s. On the other hand, several minority groups, including blacks, Hispanics, and Native Americans, have low levels of participation in science and engineering occupations both compared with other groups and compared with their proportion of the general working-age population (table 3-25). Both blacks and Hispanics also have low levels of participation in S&E relative to their proportion in the general population with a college degree. The composition of the S&E workforce across these groups has been a concern of policymakers who are interested in the development and utilization of human capital to maintain the United States' global competitiveness in science and engineering.

In 2008, with 3.5 million workers in S&E occupations, whites made up over 70% of the country's scientists and engineers. Whites accounted for more than 50% of workers within each of the S&E occupations (see appendix table 3-13). Whites are particularly highly concentrated in areas that focus on macrophysical systems. For example, whites were a strong majority of forestry and conservation scientists (91%); earth, atmospheric, and ocean scientists (86%); and agricultural and food scientists (82%).

Asians, with 824,000 workers in S&E occupations, accounted for 17% of scientists and engineers. They are strongly concentrated in computer engineering fields, constituting 40% of computer hardware engineers, 30% of computer software engineers, and 23% of the related occupations of electrical and electronics engineering. On the other hand, Asians participate in social science occupations at much lower rates than whites. For example, Asians account for 4% of psychologists and just 3% of sociologists and anthropologists.

The social sciences are the one occupation within S&E in which the underrepresented minorities (American Indian/Alaska Natives, blacks, Hispanics, and Native Hawaiians/Pacific Islanders) outnumber Asians. Collectively, these groups account for 17% of sociologists and anthropologists, and 12% of psychologists. These minorities also account for a comparatively high share of computer support specialists (16%) and statisticians (14%). On the other hand, underrepresented minorities account for relatively few physicists and astronomers (6%). Moreover, among these minority physicists and astronomers, only one-third were born in the United States compared with the more than two-thirds of underrepresented minorities who are in other S&E occupations and were born in the United States. U.S.-born underrepresented minorities accounted for less than 2% of physicists and astronomers.

Race/Ethnicity Trends in S&E Occupations

Over the past two decades, the U.S. workforce in S&E occupations has been becoming more diverse with increasing numbers of minorities and Asians and a decreasing share of whites. In 1993, 84% of workers in S&E occupations reported their race as white. By 2008, this proportion declined to 72%. Some of this decline reflects changes to the NCSES workforce surveys that collect information on race in the S&E workforce. After 2000, respondents were able to report two or more races rather than just one. Some of those who self-reported as white in the 1990s may have instead reported a multiracial identity after 2000 if they had the option, which would have decreased the estimated numbers of whites. However, because less than 2% of S&E workers self-reported two or more races in years when the option was available, it is unlikely that this change contributed to much of the decline in the share of whites between 1993 and 2008. Most of the decline in whites was offset by growth among Asians during this period and some by growth in other groups, particularly Hispanics (table 3-26).

Age Differences Among Racial/Ethnic Groups

The age structure of different demographic groups (see table 3-24) reflects the fact that members of the different groups entered the S&E workforce in different numbers at different times. The largest demographic group, whites, is also the oldest, with a median age of 44. Almost one-third of whites were older than 50 and only 13% were age 30 or younger. Blacks (median age of 42) and Hispanics (median age of 39) are somewhat younger. Asians are even younger, with a median age of 38. The comparative youthfulness of Asians reflects the age distribution of Asians working in S&E who were born in the United States. Native-born Asians were dramatically younger than other demographic groups, including foreign-born Asians. The median age among U.S. native-born Asians working in S&E occupations was 30, and only 9% were older than 50.

Racial/Ethnic Differences Among S&E Degree Holders

Most patterns across demographic groups among workers in S&E occupations also hold for members of the workforce with a highest degree in an S&E field. Additionally, outcomes that vary by race among S&E degree holders deal with unemployment rates and level of degree attainment.

In 2008, among those whose highest degree was in an S&E field, Hispanics and blacks had the highest unemployment rate (5.2% and 5.1%, respectively), which was roughly two percentage points higher than the unemployment rate for whites (3.2%). Although whites had the lowest unemployment rate, they also had the highest labor force nonparticipation rate (17%). Because of the large numbers of whites who are out of the labor force, whites have the lowest rates of employment among S&E highest degree holders.

Among those who are employed and whose highest degree is in an S&E field, race/ethnicity groups have concentrations in different degree fields. Differences in degree fields resemble those among S&E occupations. Both blacks and Hispanics are more concentrated in the social sciences, and Asians are more concentrated in engineering and in computer and mathematical sciences. In 2008, among blacks, more than half had their highest S&E degree in the social sciences, while 46% of Hispanics did (table 3-27). For both of these groups, close to one-third had their highest S&E degrees in engineering or in the computer and mathematical sciences. Asians, on the other hand, are heavily concentrated in the computer and mathematical sciences and in engineering, with 59% having their highest degree in one of these two fields and 20% having their highest degree in the social sciences. The distribution of degree fields for whites more closely resembles that for non-Asian groups. (See appendix table 3-14 for more detailed data on S&E degrees by race and Hispanic status.) On the whole, the field differences among S&E degree holders are more pronounced than are the corresponding differences among workers in S&E occupations.

In addition to having concentrations in different fields, the demographic groups differ in the level of their highest degree. For example, among Asians with a highest degree in an S&E field, 56% have their highest degree at the bachelor's level and 12% have a doctoral degree (figure 3-31). In comparison, among both blacks and Hispanics 79% have their highest degree at the bachelor's level and 4% have a doctoral degree.

Asians whose highest degree is in an S&E field are more likely than are others to work in an S&E occupation and are more likely than are others to work in the area in which they earned their degree (see appendix table 3-12). Among blacks, only one-quarter work in an S&E occupation; among Hispanics and American Indians/Alaska Natives nearly one-third work in an S&E occupation. By comparison, more than half of Asians work in these occupations.

Race/ethnicity matters even for those with similar credentials. Some, but not all, of the high concentration of black S&E degree holders working outside of science and engineering, and the high concentration of Asian S&E degree holders working within S&E, can be explained by their different degree levels or fields. But Asians with an S&E degree have a higher propensity to work in S&E occupations than others even among individuals with similar degree levels and fields. Thus differences between Asians and blacks in the propensity among degree holders to work in S&E occupations remain even among those with the same degrees.

Salary Differentials for Women and Minorities

Women and minority groups generally receive less pay than their male and white counterparts. The median salary in 2008 among women with a highest degree in an S&E field and working full time was one-third lower than the median salary among similar men (appendix table 3-15). Salary differences between men and women are much greater among those who are not working in S&E occupations. Among those working full time in S&E occupations, women's salaries were 18% lower than men's.

Racial/ethnic salary differences were somewhat smaller than salary differences between men and women (appendix table 3-16). American Indians/Alaska Natives with a highest degree in an S&E field and working full time earned 22% less than whites; blacks earned 22% less than whites; and Hispanics earned 14% less than whites. These salary differences were generally more modest among those who worked in S&E occupations.

Overall, both salary differences between men and women and race/ethnicity salary differences remained largely unchanged in the 15-year period between 1993 and 2008.

Differences in average age, work experience, field of degree, sector of employment, and other characteristics can make direct comparison of salary and earnings statistics misleading. Statistical models can estimate the size of the wage difference between men and women, as well as the wage difference between minorities and whites when various salary-related factors are taken into account. Estimates of these differences vary somewhat depending on the assumptions that underlie the statistical model used. The remainder of this section presents estimates of the expected size of the wage difference between men and women among individuals who are similar in age, work experience, field of degree, and other relevant characteristics; data bearing on wage differences between non-Asian minorities and whites are also included. These estimates are substantively consistent with many of the other published analyses on these topics (see, for example, Xie and Shauman 2003). Without accounting for any factors except level of degree, women working full time whose highest degree is a bachelor's in an S&E field were paid salaries that were 38% lower than those of men (figure 3-32).[14] This salary difference is substantial, but it is smaller at both the master's level (28%) and at the doctoral level (24%). The salary differences for minorities relative to whites are narrower (figure 3-33). Minority salary levels are 10% lower than those of whites at the bachelor's level, 16% lower at the master's level, and 4% lower at the doctoral level. All estimated baseline differences are statistically significant.

Effects of Occupation and Experience on Salary Differences

Salaries differ across occupations. For example, in the three S&E occupations with the lowest concentrations of women—aerospace, aeronautical, or astronautical engineers; mechanical engineers; and electrical and electronics engineers—the combined median salary among men is $90,000, and among women it is $81,000. These figures are substantially higher than the combined sex-specific median salary ($65,750 for men and $54,000 for women) in the three large S&E occupations with the highest concentrations of women—psychologists; medical scientists, except practitioners; and biological scientists (see appendix table 3-15). Salary also varies by indicators of experience, including both age and years since completing education. Estimates of salary differences are made by applying controls for occupation, age, and years since completing the highest degree.[15] After controlling for these factors, the estimated wage difference between men and women narrows. However, among men and women in similar jobs and with similar levels of experience, women are still paid 16% less than men (among individuals whose highest degree is at the bachelor's level) and 9% less than men (among individuals whose highest degree is at the master's and doctoral level). Minorities with their highest degree at the bachelor's level also earn somewhat less (6%) than whites, after controlling for occupation and experience. Among those with a doctoral degree, the wage difference between minorities and whites is mostly attenuated (3%) and at the master's degree level, the difference is fully attenuated after controlling for occupation and experience. This illustrates that at higher degree levels, minorities and white degree holders in similar S&E occupations and with similar experiences receive about the same salaries.

Effects of Other Factors: Sector, Field of Degree, and Region

Salaries vary by other work-related factors beyond occupation and experience. For example, salaries differ across sector. 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. These differences are salient for understanding salary variations by sex and race/ethnicity because whites and males are more highly concentrated in the private for-profit sector. Salaries also differ across regions. For example, at $86,000, the Pacific census division had the highest median salaries for scientists and engineers among the nine U.S. census divisions and the west-north-central, at $75,000, had the lowest. Almost one-quarter (23%) of U.S.-born underrepresented minorities worked in the Pacific division compared with 15% of whites, whereas whites had a higher concentration in the west-north-central (9%) than underrepresented minorities (4%).

Salaries also vary by degree field. Salaries among those with degrees in engineering, the physical sciences, and in computer and mathematical sciences are higher than salaries among those with degrees in the environmental and life sciences, and among those with degrees in the social sciences. Degree areas with lower salaries also have higher concentrations of women and minorities.

However, taking these factors into account[16] in addition to occupation and experience results in only marginal changes in the estimated salary differences between men and women compared with estimates generated accounting for occupation and experience alone. Women who are similar to men along all seven of these factors receive salaries that are 13% (among bachelor's degree holders) to 8% (among master's degree and doctoral degree holders) lower than their male counterparts. The salary difference between minorities and whites fully attenuates when all seven factors are simultaneously controlled.

Effects of Family on Salary Differences

The family roles of wife and mother are associated with lower salaries for women. In contrast, the roles of husband and father are associated with higher salaries among men. To evaluate the effects of family status on wage differences between men and women, these differences are estimated separately for the set of workers in science and engineering occupations who are unmarried and without young children, who are married and without young children, and who are married and with young children. Each estimate is made accounting for occupation, age, time since degree, employment sector, field of degree, region, and parents' educational attainment, as described above. The analysis presented in figure 3-34 considers a household to include young children if a child age 12 or younger was present.[17]

Among full-time workers with a highest degree in an S&E field who are both unmarried and childless, men and women tend to be paid similar salaries. At the bachelor's level, the estimated salary difference is 3% among men and women who are similar in occupation, age, experience, work sector, degree field, region, and parents' education (figure 3-34). At the master's and doctoral levels, estimated salary differences between men and women among the unmarried and childless are statistically insignificant. The presence or absence of children under age 12 does not consistently affect the size of salary differences between men and women beyond what would be expected considering other factors.[18]

S&E Immigrants

The foreign born constitute a considerable proportion of workers in science and engineering occupations, and both the number and share of foreign-born workers have been increasing. However, immigration of scientists and engineers to the United States has declined during the recent economic downturn. Most indicators presented in this section apply to all foreign born, despite the fact that the foreign born is a broad category comprising long-term U.S. residents with strong roots in the United States as well as recent immigrants who compete in global job markets or whose main social ties are in their countries of origin.

Several sources yield broadly consistent estimates of U.S. reliance on foreign-born scientists and engineers. Table 3-28 shows upward trends in the percentage of foreign-born individuals in U.S. S&E occupations over the first decade of the century. The share of nonacademic scientists and engineers who are foreign born rose from 22% in 2000 to 25% in 2009, although some evidence suggests that the rate of growth slowed in the last years of the decade.

The similarity in the estimates from SESTAT and the U.S. Census Bureau American Community Survey (ACS) is noteworthy because the two surveys differ methodologically. SESTAT surveys include only individuals who were counted in the most recent Decennial Censuses or who received a U.S. S&E degree, thereby excluding recently arrived foreign-born and foreign-educated scientists and engineers. The potential for an undercount of the foreign born is smallest in the earliest portion of the decade—the closer in time to the Decennial Census—and increases over the course of the decade.[19] The ACS, on the other hand, draws a new sample of the U.S. residential population every year. However, ACS occupation coding is less precise, and the ACS does not distinguish postsecondary teachers in science and engineering fields from other postsecondary teachers. The similarity in the estimates from these surveys, despite their contrasting limitations, suggests that the overall picture the surveys provide is broadly accurate.

Characteristics of the Foreign Born

The foreign born in S&E occupations tend to have higher levels of education than the U.S. native born. In most S&E occupations, the higher the degree level, the greater the proportion of the workforce who are foreign born (appendix table 3-17). This relationship is weakest among social scientists and strongest among computer and mathematical scientists and engineers. In 2003, at the bachelor's degree level, the proportion of foreign-born individuals within occupational areas ranged between 10% (social scientists) and 19% (computer and mathematical scientists). However, at the doctoral degree level, about half of the workers in computer and mathematical sciences and in engineering were foreign born.

In 2003, more than half (55%) of foreign born in the United States with a highest degree in an S&E field came from Asian countries. Just over one-fifth were born in Europe. North America (Canada), Central America, the Caribbean, South America, and Africa each supply roughly equal numbers (each accounting for from 4% to 5% of the foreign born). The leading country of origin among immigrant S&E workers in the United States is India, which accounted for 16% of the foreign born. China (with 11%) is the second leading country. Source countries for the 276,000 foreign-born holders of S&E doctorates are somewhat more concentrated, with China providing 22% and India 14% (figure 3-35 and appendix table 3-18).

Source of Education

The majority of foreign-born scientists and engineers in the United States came to the United States before completing their higher education, but a substantial number came to the United States after receiving their university training abroad. Although almost half of the foreign-born, university-educated individuals working in the United States have a degree from a foreign university, two-thirds earned their highest degree from a U.S. educational institution (see appendix table 3-18). Among the foreign born with a doctoral degree, just over two-thirds received this degree from a U.S. institution, although nearly 80% have at least one degree from a foreign institution.

New Foreign-Born Workers

The number and share of foreign-born S&E workers have been rising, but the volume of new foreign workers entering U.S. S&E occupations has shown signs of decline during the recent economic downturn. One indicator of new foreign-born S&E workers joining the U.S. workforce is the number of temporary work visas issued by the U.S. government in visa classes for high-skilled workers. A second indicator is the rate at which foreign-born recipients of U.S. doctoral degrees remain in the United States after earning their degree ('stay-rates').

Temporary Visas

The number of temporary work visas issued for high-skill workers provides an indication of the volume of immigration of these workers. However, for all types of temporary work visas, the actual number of individuals using them is less than the number issued. For example, some individuals may have job offers from employers in more than one country and may choose not to foreclose any options until a visa is certain.

The largest classes of these temporary visas declined during the recent economic downturn, after several years of growth (figure 3-36). Data for 2010, however, suggest that this period of decline may be short-lived. The previous period of decline in the use of these visas occurred during the more mild recession in the earlier part of the decade, and these declines were unevenly experienced across visa categories.

H-1B temporary work visas account for a larger number of high-skill workers than other visa classes. This visa is issued to individuals who seek temporary entry into the United States in a specialty occupation that requires the skills of a professional. It is issued for up to 3 years with the possibility of an extension to 6 years. In 2010, the United States issued more than 118,000 H-1B visas, down almost 25% from the nearly 155,000 issued in 2007.

Similarly, many fewer J-1 exchange visas—visas issued for limited periods of study, research, or teaching—were issued in 2010 than in 2007. For L-1 visas, which support intercompany transfers, the number was 12% lower in 2010 than in 2007. The smaller, more specialized visa programs for high-skilled workers also fell slightly in 2010. These visa classes include O-1 (a person of outstanding ability), O-2 (an assistant to an O-1, sometimes a postdoc), TN (college-degreed citizens of Canada and Mexico), and E-3 (college-degreed citizens of Australia).

Characteristics of H-1B Visa Recipients

The H-1B visa, which is the most common visa for new foreign entrants into the U.S. S&E workforce, is not issued exclusively for scientists and engineers. Other professional workers who use an H-1B visa include those in administrative occupations, legal occupations, and cultural occupations (such as artists and entertainers). However, because the U.S. Citizenship and Immigration Services do not classify occupations with the same taxonomy used by the National Science Foundation, precise counts of H-1B visas issued to scientists and engineers cannot be obtained. Nevertheless, it is safe to say that the bulk of H-1B visa recipients work in S&E or S&E-related occupations (appendix table 3-19). In 2009, workers in computer-related occupations were the most common recipients of H-1B visas, accounting for 35% of H-1B visas issued. The total number of newly initiated H-1B visas for workers in computer-related fields declined by nearly half from 2008 to 2009 while the share of total recipients who worked in these fields declined from 50% to 35%. Despite this drop, the proportion of H-1B recipients who worked in computer sciences was considerably higher than it was early in the decade. For example, in 2002, only 25% of these visa recipients worked in computer-related fields.

H-1B visa recipients tend to possess advanced degrees. In FY 2009, 58% of new H-1B visa recipients had an advanced degree, including 40% with master's degrees, 6% with professional degrees, and 13% with doctorates. This degree distribution differs by occupation, with 83% of mathematical and physical scientists holding advanced degrees (44% with doctorates). Among life scientists, 87% hold advanced degrees (61% with doctorates).

Almost half of recent H-1B visa recipients were from India (39%) or China (10%). Among doctorate holders, 29% were from China and another 16% from India (figure 3-37). Altogether, Asian citizens made up nearly two-thirds of all H-1B visa recipients with a doctoral degree.[20]

Table 3-29 shows salaries paid to new recipients of H-1B temporary work visas by occupation group and level of degree. These starting salaries, taken from final visa application forms sent to U.S. Citizenship and Immigration Services, are different from—and generally higher than—H-1B salaries that firms report on their applications to the Department of Labor, which are filed much earlier in the H-1B process. The relatively low average salaries for doctorate holders in the life sciences may reflect the common use of H-1B visas to hire individuals for relatively low-paying postdoc fellowships.

Short-Term Stay Rates for U.S. Doctorate Recipients

Among doctoral recipients, the period immediately after earning the doctoral degree is a pivotal point at which long-term career trajectories may be set. Foreign doctoral recipients who remain in the United States may set themselves on a pathway toward long-term residency.

At time of award, foreign students who receive doctoral degrees from U.S. universities report whether they intend to stay in the United States and whether they have a firm offer (either a postdoc or employment opportunity) to stay in the United States.[21] These responses provide estimates of short-term stay rates.

Most foreign U.S. doctorate recipients plan to stay in the United States after graduation. At the time of doctorate receipt, three-quarters of foreign recipients of U.S. S&E doctorates, including those on both temporary and permanent visas, plan to stay in the United States, and about half have either accepted an offer of postdoc study or employment or are continuing employment in the United States (figure 3-38).[22] Through the 1980s, about half of foreign students who earned S&E degrees at U.S. universities reported that they planned to stay in the United States after graduation, and about one-third said they had firm offers for postdoc study or employment (NSB 1998). In the 1990s, however, these percentages increased substantially. Thus, the proportion of foreign S&E doctoral degree recipients reporting plans to stay in the United States rose to 72% in the 1998–2001 period and to 77% in the 2006–09 period (appendix table 3-20). In 2009, both the percentage who reported plans to stay in the United States and those with firm offers to stay declined modestly from 2008. The number of foreign doctoral degree recipients also declined in 2009, making the drop from 2008 to 2009 in numbers of foreign-born doctoral recipients with plans to stay in the United States somewhat more pronounced, with 6% fewer foreign-born doctoral recipients reporting plans to stay in the United States.

Overall S&E short-term stay rates reflect the high short-term stay rates in computer and mathematical sciences, the biological and related sciences, the physical sciences, and engineering. Between 2006 and 2009, the short-term stay rate in each of these four fields was about 80%, as measured by reports of intentions to stay in the United States. However, the short-term stay rate for foreign doctoral recipients in the social sciences and in health fields was considerably lower.

Stay rates vary by place of origin. In the period 2006–09, 89% of U.S. S&E doctoral recipients from China and from India reported plans to stay in the United States, and close to 60% reported accepting firm offers for employment or postdoc research in the United States. Doctorate recipients from Japan, South Korea, and Taiwan were less likely than those from China and India to stay in the United States (figure 3-39). Close to half of U.S. S&E doctoral degree recipients from Europe had firm plans to stay after graduation. In North America, the percentage of 2006–09 doctoral degree students who had definite plans to stay in the United States was higher for those from Canada than those from Mexico (see appendix table 3-20).

Between 2002–05 and 2006–09, the percentage of U.S. S&E doctoral degree recipients from the two top countries of origin (China and India) who were reporting definite plans to stay in the United States declined. Other countries, however, experienced sharp increases in short-term stay rates among S&E doctoral degree recipients in the United States, including Indonesia, New Zealand, Mexico, and Colombia.

Long-Term Stay Rates

The rate at which foreign recipients of U.S. doctoral degrees who stayed in the United States immediately after they received their degree continue to remain in the United States over longer durations can also be observed.[23] Recent trends in long-term stay rates show that within cohorts, long-term stay rates are similar to short-term rates. This similarity is particularly evident for the cohort of foreign S&E doctoral recipients who earned their degrees in 1993 (figure 3-40). Two years after receiving the doctoral degree, 53% of these foreign doctorates who were temporary residents when they earned their degree remained in the United States. By 2009, 52% remained, with little variation along the way. More recent cohorts have had higher short-term stay rates, but these stay rates have declined over time. The cohort of degree recipients who earned their doctorates in 1999 had a stay rate after 2 years of 68%. After 10 years, this rate declined by 7 percentage points, but the rate of decline gradually attenuated. The cohort of foreign S&E doctoral degree recipients of 2004 had a 2-year stay rate of 66%, which declined to 62% by 2009 (figure 3-40; Finn 2012, forthcoming).

The stability of stay rates over time applies whether or not these rates are calculated for foreign doctoral recipients from U.S. institutions who received their doctoral degree while on a temporary visa status or for those who held either a temporary or permanent visa. Temporary visa holders make up the largest share of foreign S&E doctoral degree recipients. They also have lower stay rates than do permanent residents. For example, among foreign S&E doctoral degree recipients from the 1993 cohort, those who were permanent residents at the time they earned their degree had stay rates that were 24 percentage points higher than those with temporary visas. This difference persisted through 2009. Among more recent cohorts, the difference in stay rates between permanent and temporary residents was initially much smaller, but increased rapidly over the 5 years after receipt of the doctorate.

Because of the persistence of stay rates over time, factors that are associated with the level of short-term stay rates are similarly associated with the level of longer-term stay rates. For example, countries with the highest levels of short-term stay rates (e.g., China and India) are among the countries with the highest long-term stay rates. Similarly, academic fields that have the highest short-term stay rates (e.g., the physical sciences) also have the highest long-term stay rates, and the field with the lowest short-term stay rates, the social sciences, has the lowest long-term stay rates.

Some evidence suggests that stay rates may vary for doctorate recipients from graduate programs of different quality based on ratings of faculty by the publication U.S. News and World Report and on separate ratings by the National Research Council (Finn 2009). Doctorate recipients from the graduate programs designated among the top 25 were somewhat less likely to remain in the United States than were graduates of other programs (see table 3-30). The difference in 1-year stay rates was 2 percentage points: 69% of those from the top-rated programs and 71% of other doctorate recipients remained in the United States 1 year after receiving their degrees. By 5 years after receiving their degree, the two groups showed differences that rose to 5 percentage points, with stay rates of 59% and 64%, respectively.

Age and Retirement

The baby boom generation—the unusually large cohort born between 1946 and 1964 (with birth rates in the United States peaking in 1957)—affected the age structure of the S&E labor force in much the same way it affected the general labor force. Thus, in the early 1990s, this bulge produced a relatively large concentration of S&E workers in their late 20s to mid-40s contributing to a comparatively youthful S&E workforce. By 2008, these cohorts had aged into their early 40s to early 60s, with the oldest nearing traditional retirement ages. One indicator of the aging of the S&E workforce is the increasing percentage of individuals in this workforce above age 50 (as seen in figure 3-41). In 2008, 27% of individuals with S&E degrees and in S&E occupations were in that age group, whereas 15 years earlier just 18% were in that age group.

Another indication of the aging of the S&E labor force is the increase over time of the median age of individuals working in S&E occupations. From 1993 to 2008, the median age rose by 4 years, from 37 to 41 years of age. The median age of workers with a highest degree at the bachelor's level rose by 5 years (from 35 to 40), at the master's level by 3 years (from 39 to 42), and at the doctoral level by 3 years (from 44 to 47).

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).

Age Differences Among Occupations

Individuals with S&E degrees who work in S&E occupations are younger than individuals with S&E degrees who work in S&E-related occupations. They also are younger than those whose jobs are not in, nor related to, S&E. Figure 3-42 shows, for 2008, age distributions for S&E-degree holders by highest degree level and broad occupational area. Age differences across broad occupational areas are more pronounced at higher degree levels. Among those whose highest S&E degree is at the master's level, the median age of workers in S&E occupations was 42; for workers in S&E-related occupations it was 47; for workers in jobs not in nor related to S&E occupations it was 49. Among those whose highest S&E degree is at the doctoral level, the median age of workers in S&E occupations was 47 compared with 50 for workers in S&E-related occupations and 53 for workers in jobs not in nor related to S&E. The flow of workers out of S&E occupations into other occupations compared with the reverse flow from other occupations to S&E occupations contributes to much of the differences in age distributions across broad occupational areas. For example, among workers in S&E occupations who were observed in 2003, 16% were no longer in such occupations in 2006. On the other hand, only 5% of those workers in other occupations in 2003 were in S&E occupations in 2006. Among the S&E workers who moved into other occupations, one-third (approximately 200,000 workers) went into management positions, many of which involve supervising S&E workers.

Age Differences Among S&E Degree Fields

In 2008, the median age among those in the labor force with any degree in S&E was 43. Degree holders in different areas varied in their ages. Degree holders in the physical sciences were comparatively old with a median age of 47 and 38% of the field's workers over age 50 (figure 3-43). Degree holders in computer and mathematical sciences were relatively young, with a median age of 42 and only 22% over age 50. Within degree areas, specific fields differed considerably in the ages of their workers. For example, within engineering the youngest degree holders were in bioengineering and biomedical engineering, with a median age of 34 and with 39% younger than age 30 (see appendix table 3-21). On the other hand, more than 40% of the workers in metallurgical engineering and mining and mineral engineering were older than 50.

Leaving the Labor Force and Retirement

The increasing share of the S&E labor force over age 50 makes retirement patterns among S&E workers more important in terms of how they will affect the supply of these workers. Recent patterns of labor force exit and work reduction among the older members of the workforce suggest that by age 55 rates of participation in the S&E workforce begin to decline and are markedly reduced by the time workers reach their mid-60s. One indication of the relationship between age and the level of labor force participation is illustrated by figure 3-44, which shows full-time work rates among older S&E degree holders by highest level of education. In 2008, at age 55, 78% of those whose highest degree was at the bachelor's level, 75% of those whose highest degree was at the master's level, and 89% of those whose highest degree was at the doctoral level worked full time. However, at all degree levels, full-time labor force participation rates decline quickly as S&E workers age into their late 50s. By age 61, more than half of S&E bachelor's degree holders are not working full time. Among those whose highest degree is at the master's level, this milestone is reached at age 62. For S&E doctoral degree holders, half are not working full time by age 64. After age 65, no more than one-quarter of the workforce with a highest degree at the master's or bachelor's level worked full time. Among those with a doctoral degree, this proportion is reached at age 71.

Another indicator of the relationship between age and labor force participation is the proportion of S&E degree holders who reported that they were out of the labor market. In 2008, at age 55, 12% of those whose highest degree is at the bachelor's level, 7% of those whose highest degree is at the master's level, and 5% of those whose highest degree was at the doctoral level were out of the labor force. By the early 60s, the proportion of people who are out of the labor force takes a sharp turn upwards, and by age 65 about half of those whose highest degree is at the master's level and half of those whose highest degree is at the bachelor's level report that they are neither working nor looking for work. Among those with a doctoral degree, more than half report neither working nor looking for work at age 68.

Table 3-31 shows the rates at which holders of U.S. S&E doctorates left full-time employment, by sector of employment, between April 2006 and October 2008. Rates of leaving full-time employment for S&E doctorate holders were higher for those working in the private sector than those employed in education or government, although in the oldest group this sector difference largely disappears.

Between 1993 and 2008, increasing percentages of SESTAT respondents in their 60s reported that they were still in the labor force. Whereas 59% of S&E degree holders between the ages of 60 and 64 were employed in 1993, the comparable percentage rose to 66% in 2006. For S&E degree holders between ages 65 and 69, the increase was larger, rising from 32% in 1993 to 44% in 2006. After peaking in 2006, rates of employment among workers in their 60s declined slightly in 2008, but remained above rates prior to 2006. Other indicators, including full-time employment rates and retirement rates, show similar patterns, as do comparisons restricted to workers with similar highest degree levels and degrees in similar fields. In recent years, labor force participation has also risen slightly among S&E degree holders in their early 70s, but has not changed among those in their late 50s.


[14] This estimate differs slightly from the observed median difference in salary by sex because the former addresses mean differences and the latter addresses median differences. The former is influenced by extreme cases and outliers, and the latter is not.
[15] Occupation, age, and years since completion of education are each controlled for as a random effect. SESTAT respondents working in science and engineering have been classified into 62 distinct occupations. Age is observed in one of eleven 5-year brackets. Years of experience are observed in one of twelve 5-year brackets.
[16] Occupational sector, region, field of degree, and parents' education are each controlled for as a random effect. Employers are classified into one of seven sectors: 4-year colleges and universities, 2-year colleges, for-profit private sector, nonprofit private sector, self-employment, federal government, and state and local government. Regions are classified into the nine U.S. census divisions. Field of degree is observed in 1 of 142 distinct degree fields among individuals whose highest degree is at the bachelor's level, and within 123 distinct degree fields among individuals whose highest degree is at the doctoral level. Parents' education measures the highest level of education completed by either parent and is observed in one of eight categories.
[17] The analysis was repeated with different age cut-points defining young children. Results did not change substantially when this age limit was adjusted (from ages 0–18 to ages 0–6), indicating that the finding in the text is not substantively sensitive to where this cut-point is set.
[18] Among married workers with children younger than age 12, the estimated salary differences between men and women are generally similar in magnitude to the estimates for all scientists and engineers. For example, among workers whose highest degree is a bachelor's in an S&E field, the estimated salary difference by sex is 13% among all workers and is also 13% among workers who are married and with children younger than age 12. At the doctoral level, the estimated 8% salary difference by sex applies to all workers and to workers who are married with children. Only at the master's degree level is the estimated salary difference between men and women among the married with children larger (at 15%) than the difference among all workers (7%).
[19] In the future, however, the largest component of SESTAT, the National Survey of College Graduates, will be refreshed on a biennial basis using respondents from the ACS, and so the undercount of recent foreign arrivals will be minimized.
[20] This includes East Asians, South Asians, and Southeast Asians, but excludes individuals from countries in the Middle East and from the former Soviet Republics.
[21] This question is part of the Survey of Earned Doctorates, which is administered to all recipients of U.S. doctoral degrees.
[22] The growth in the number of doctoral students from China accounts for much of the rapid increase in foreign recipients of doctoral degrees from the early 1980s through 1996. During this period, the annual count of Chinese recipients of doctoral degrees rose from fewer than 10 to more than 3,000 (from 0.1% to 27.4% of all foreign doctoral degree recipients). The decline in foreign doctoral degree awards following 1996 also is partially, but not fully, accounted for by changes in the numbers of Chinese doctoral degree recipients. One contributing factor to the decline in 1996 was the Chinese Student Protection Act of 1992.
[23] Long-term stay rates are observed by annually calculating the ratio of the number of noncitizen Survey of Earned Doctorate respondents who made Social Security contributions to the number of noncitizen Survey of Earned Doctorate respondents.