Women Scientists and Engineers
- Employment and Unemployment
- Sector of Employment
Women are 22 percent of the science and engineering labor force as a whole (see figure 5-1) and were 20 percent of doctoral scientists and engineers in the United States in 1993, compared with 19 percent in 1991. 
See appendix tables 5-1, 5-2, and 5-3.
Within science and engineering, women are more highly represented in some fields than in others. Women are more than half of sociologists and psychologists but are only 9 percent of physicists and 8 percent of engineers. (See appendix table 5-1.) Doctoral women scientists and engineers are likewise more heavily represented in some fields than in others. For example, women are 41 percent of doctoral psychologists, and 28 percent of biologists but only 4 percent of engineers. (See figure 5-2.)
See appendix table 5-5.
In many fields, women scientists and engineers are much more likely than men to have the bachelor's degree as their highest degree. Women are 32 percent of bachelor's computer/mathematics scientists but only 18 percent of doctoral computer/mathematics scientists. (See appendix table 5-1.) Because of these differences in highest degree, the science and engineering work done by women is often very different from that done by men. For example, in the biological sciences, women are 47 percent of the bachelor's biological scientists and only 29 percent of the doctoral biological scientists. (See appendix table 5-1.) Biological scientists with bachelor's degrees may have as their primary activity testing and inspection or technical sales or service, or they may be biological technicians, medical laboratory technologists, or research assistants. Biological scientists with doctoral degrees typically teach in universities, perform independent research, or are managers or administrators in industry. 
Employment and Unemployment
Bachelor's and Master's Scientists and Engineers
Recent men and women bachelor's science and engineering graduates are similar in their pursuit of postgraduation education but differ in employment status. About 30 percent of new bachelor's graduates do not immediately seek employment. Instead, they pursue graduate study either full time or part time. (See figure 5-3.) In 1993, women and men 1992 science and engineering graduates were about as likely to be enrolled in graduate school (32 percent of women versus 29 percent of men). (See appendix table 5-6.)
See appendix table 5-6.
Recent men and women bachelor's graduates differ more in postgraduation employment status than they do in postgraduation education. Men bachelor's science and engineering graduates are more likely to be in the labor force, to be employed full time, and to be employed in their field than are women. (See figure 5-4.) Women are more likely than men to be out of the labor force, to be employed part time, and to be employed outside their field. Women are 44 percent of the 1992 bachelor's science and engineering graduates but are 58 percent of those out of the labor force (i.e., not employed and not seeking employment), 54 percent of those employed part time, and 47 percent of those employed full time outside their field. (See appendix table 5-6.)
See appendix table 5-6.
Some of these differences are due to family-related reasons, often demands of a spouse's job or presence of children. Among recent bachelor's graduates, 29 percent of women but only 1 percent of the men who are not employed cited family responsibilities as the reason for not working. (See appendix table 5-7.)
Field differences contribute to some of these differences in employment status as well. Undergraduate education in science and engineering is not necessarily preparation solely for science and engineering employment. Science and engineering education at the undergraduate level is broadly applicable in a number of fields outside science and engineering.
Among employed recent science and engineering bachelor's graduates, women are less likely than men to be employed in science and engineering occupations. Only 18 percent of the employed new women graduates compared with 35 percent of the new men graduates are employed in science and engineering. (See appendix table 5-8.) Those who are not employed in science and engineering occupations are, for the most part, in related occupations, such as clinical psychology, social work, management, secondary education,  and sales and marketing. (See figure 5-5.) Because they are more likely than men to earn degrees in the social sciences, women are more likely than men to be employed in social services and related occupations and, because of family concerns, cultural norms, or personal preference, are more likely than men to be employed in secondary education.
See appendix table 5-8.
Part of the reason women bachelor's science and engineering graduates are less likely than men to be employed in science and engineering occupations is that women are not highly represented in fields in which a bachelor's degree is sufficient for employment within the field. Engineering and computer science, fields in which women are not highly represented, typically provide "professional" employment with bachelor's degrees. Thus, new bachelor's graduates in these fields are likely to find employment in their field: 72 percent of 1992 bachelor's computer science graduates and 65 percent of new bachelor's engineering graduates found full-time employment in their field.
Other fields typically require graduate education for "professional" employment in the field. New bachelor's graduates in these fields are least likely to be employed within their field. Life sciences and social sciences, fields in which women are highly represented, are two such fields: only 37 percent of 1992 bachelor's social science graduates and 32 percent of 1992 bachelor's life science graduates found full-time employment in their field.
Unemployment rates of men and women recent bachelor's graduates do not differ greatly: 4.1 percent of women and 4.7 percent of the men 1992 bachelor's science and engineering graduates were unemployed in April 1993. (See appendix table 5-6.)
Doctoral Scientists and Engineers
The overall labor force participation rates of doctoral men and women scientists and engineers are similar-about 92 percent of both men and women are in the labor force. The labor force participation rates of men and women who received their doctorate in similar time periods are quite different, however. Within degree cohorts, men have higher labor force participation rates than women. For example, among 1980-1984 graduates, the labor force participation rate for men is 99.1 percent; for women, it is 93.8 percent. (See appendix table 5-9.) Because a higher fraction of men than women are in the earlier degree cohorts (e.g., those who received degrees before 1960) and those in earlier degree cohorts have lower labor force participation rates, largely due to retirements, men's overall participation rate averages out to about the same as women's.
Among doctoral scientists and engineers, 12 percent of women and 4 percent of men are employed part time. (See appendix table 5-10.) Women who are employed part time are far more likely than men to cite family responsibilities as the reason. (See appendix table 5-11.) About half of the doctoral women working part time and about 5 percent of the men cited family responsibilities as the reason for working part time. Women with children under age 18 are more likely than men with or without children and women without children to work part time or to be unemployed. (See appendix table 5-12.)
Women and men who have children face the problem of trying to balance work and family. Twenty-one percent of doctoral women scientists and engineers with children under 18, but only 2 percent of comparable men, are employed part time. Both men and women face the problem of balancing work and family when employers demand primary commitment to work. Even companies with family-friendly programs frequently discourage their use. 
New doctoral scientists and engineers are more likely than bachelor's scientists and engineers to find employment in their field. Among full-time employed doctoral scientists and engineers, 93 percent are employed in their field, compared with 70 percent of full-time employed bachelor's scientists and engineers. (See appendix tables 5-6 and 5-10.) Doctoral women who are employed full time are as likely as men to be in jobs related to their degree.
Family status influences exit rates out of science and engineering employment. Married scientists and engineers and those with children are more likely to leave science and engineering employment than those who are not married and do not have children.  Within each family status category, however, differences between men and women remain. Single women are more likely than single men to leave science and engineering employment. Married women without children are more likely than married men without children to leave science and engineering employment, and women with children are more likely than men with children to leave science and engineering employment.
Women doctoral scientists and engineers are more likely than men to be unemployed, although the difference is small. The unemployment rate  for doctoral women in 1993 was 1.8 percent; for men it was 1.6 percent.  (See figure 5-6.) Within fields, the differences in unemployment rates are larger, especially in the fields that have fewer women. For example, among physical scientists, the unemployment rate for women is 3.2 percent compared with a rate of 2.0 percent for men. (See appendix table 5-13.) Among engineers, the unemployment rate for women is 2.4 percent compared with a rate of 1.7 percent for men. Among social scientists, on the other hand, the unemployment rates are more nearly equal-1.4 percent for women and 1.5 percent for men.
See appendix table 5-13.
Sector of Employment
Bachelor's and master's scientists and engineers are employed predominantly in business or industry. Seventy-two percent of bachelor's scientists and engineers, and 56 percent of master's scientists and engineers are employed in this sector. (See appendix tables 5-14 and 5-15.) Doctoral scientists and engineers, on the other hand, are employed in diverse sectors: 45 percent are employed in universities or 4-year colleges, 30 percent are employed in business or industry, 10 percent are employed in government, and 15 percent are employed elsewhere. (See appendix table 5-16.)
Among bachelor's and master's scientists and engineers, women, minorities, and persons with disabilities are less likely than scientists and engineers as a whole to be employed in business or industry and are more likely to be employed in educational institutions. For example, among master's scientists and engineers, 63 percent of men and 39 percent of women are employed in business or industry and 16 percent of men and 32 percent of women are employed in educational institutions. (See appendix table 5-15.)
Among doctoral scientists and engineers, women are also less likely than men to be employed by private for-profit employers and are more likely than men to be employed in colleges and universities or to be self-employed. (See figure 5-10.) These differences in sector are mostly related to differences in field of degree. (See appendix table 5-17.) For example, women are less likely than men to be engineers or physical scientists, who tend to be employed by private for-profit employers. Forty-four percent of doctoral physical scientists and 53 percent of doctoral engineers are employed in business or industry, compared with 30 percent of all scientists and engineers. Within fields, women are about as likely as men to choose industrial employment, although some evidence indicates that women leave industrial employment at a greater rate than men.  The climate in industry may be perceived as less favorable to women for a number of reasons including recruitment and hiring practices, a corporate culture hostile to women, sexual harassment, lack of opportunities for career development and critical developmental assignments, failure to accommodate work-family issues, lack of mentoring, and lack of access to informal networks of communication. 
See appendix table 5-16.
Women's greater tendency to be self-employed is also related to field of degree. For example, women are more likely than men to be psychologists, and psychologists are more likely than other scientists and engineers to be self-employed. Twenty-two percent of doctoral psychologists are self-employed, as opposed to only 6 percent of all scientists and engineers. (See appendix table 5-17.)
The employment characteristics of women in colleges and universities are quite different from those of men. Women faculty differ from men in terms of teaching field, type of school, full-time or part-time employment, contract length, primary work activity, research productivity, rank, and tenure.
The fields in which men and women faculty teach differ. Women faculty as a whole are less likely than men to be science and engineering faculty. Women are 44 percent of faculty in non-science-and-engineering fields but only 24 percent of science and engineering faculty. (See appendix table 5-18.) Within science and engineering, women faculty are a relatively small fraction of physical science and engineering faculty and are more highly represented among mathematics and psychology faculty. Women are 43 percent of psychology faculty and 31 percent of mathematics faculty but only 14 percent of physical science and 6 percent of engineering faculty.
The types of schools in which men and women teach differ. Women science and engineering faculty are far less likely than men faculty members to be employed in research universities and are far more likely to be employed in public 2-year schools. (See figure 5-11.) Differences in type of school are related to faculty employment status. Women science and engineering faculty are much more likely than men to teach part time (40 percent versus 25 percent). (See appendix table 5-19.) Two-year schools are much more likely than 4-year schools to hire part-time faculty. More than half of faculty, regardless of sex, who work in 2-year schools work part time. (See appendix table 5-21.)
See appendix table 5-19.
Women are also more likely than men to have fixed-term contracts. Fifty-four percent of women science and engineering faculty are on a one-term or 1-year contract, compared to 34 percent of men. (See appendix table 5-20.) Some evidence indicates that such contracts are becoming more prevalent. Over the last 5 years, colleges and universities have moved toward replacing tenured or tenure-track positions with fixed-term contracts. 
The differences among men and women faculty in type of schools and employment status are partly related to the highest degree obtained. Fewer women than men science and engineering faculty have a PhD degree. A far higher proportion of women (42 percent) than men (24 percent) faculty have a master's degree as their highest degree. (See appendix table 5-22.)
Partly because of the types of schools in which they are employed, women science and engineering faculty are more likely than men to be involved primarily in teaching. (See appendix table 5-23.) Not only do they spend more time teaching than men, they also are more likely than men to report they prefer teaching to research. Within school types, men and women faculty are more nearly the same in the amount of time spent in teaching or research and in the preferred amount of time spent in teaching or research.
Women science and engineering faculty also do less research than men faculty. Women are less likely than men to be engaged in funded research, to be a principal investigator or co-principal investigator (see appendix table 5-24), or to have published books or articles in the previous 2 years (see appendix table 5-25). These differences remain even within research universities and among all age groups.
Among full-time science and engineering faculty, women are less likely to chair departments, are less likely to reach the highest academic ranks, and are less likely to be tenured than men. Eleven percent of women but 14 percent of full-time men science and engineering faculty chair departments. (See appendix table 5-26.)
Women scientists and engineers hold fewer high-ranked positions in colleges and universities than men. Women are less likely than men to be full professors and are more likely than men to be assistant professors or instructors. (See figure 5-12.) Part of this difference in rank can be explained by age differences, but differences in rank remain even after controlling for age. Among those who received their doctorate 13 or more years ago, 72 percent of men but only 55 percent of women are full professors. (See appendix table 5-27.)
See appendix table 5-27.
Women are also less likely than men to be tenured or to be on a tenure track. Forty-three percent of full-time employed women science and engineering faculty are tenured, compared with 67 percent of men. (See figure 5-13.) As was the case with rank, some of the differences in tenure may be attributable to differences in age.
See appendix table 5-28.
As noted earlier, bachelor's and master's scientists and engineers are employed primarily in business or industry, and women scientists and engineers are less likely than men to be employed in this sector. The type of work women scientists and engineers do also differs from that done by men. For example, 40 percent of bachelor's-level women but only 26 percent of bachelor's-level men report computer applications as their primary work activity. Thirteen percent of master's-level men and 9 percent of master's-level women are managers. (See appendix table 5-29.) Age differences largely explain differences in management status. Among bachelor's scientists and engineers between the ages of 30 and 39, roughly equal proportions of men and women are managers. Differences in field also have a lot to do with differences in primary work activities. For example, men are more likely than women to be engineers and are thus more likely to be engaged in development, design of equipment, and production.
Among doctoral scientists and engineers, nonacademic employment is more prevalent than academic employment in some fields, for example, chemistry and engineering. Women are less likely than men to be employed in these fields and are less likely than men to be employed in nonacademic settings.
Within business or industry, women doctoral scientists and engineers are less likely than men to be in management. (See figure 5-14.) Twenty-five percent of doctoral men scientists and engineers and 21 percent of doctoral women scientists and engineers are in management. As was the case with bachelor's- and master's-level scientists and engineers, this difference is largely attributable to differences in age. Among employed industrial scientists and engineers who received doctoral degrees since 1985, 10 percent of men and 13 percent of women are managers. Among those who received degrees between 1970 and 1979, 32 percent of both women and men are managers. (See appendix table 5-30.)
See appendix table 5-30.
Bachelor's and Master's Salaries
The 1993 median starting salary for recent women bachelor's science and engineering graduates was lower than that for men overall, largely because of differences in occupational field. Women are less likely than men to be computer/math scientists or engineers, who earn relatively high salaries. They are more likely than men to be social or life scientists, who earn relatively low salaries. Within fields, the median starting salaries for men and women were more nearly the same. (See text table 5-1.) For example, in engineering, the median salary for men was $33,500 and for women was $33,600. The starting salaries of men and women in computer and mathematical sciences, physical sciences, and sales and marketing were very similar.
Among more experienced bachelor's and master's scientists and engineers, the gap between men's and women's salaries is larger. (See appendix table 5-31.) As was the case for starting salaries, some of the differences in salary are due to differences in field. Salaries are highest in mathematical/computer science and engineering, fields in which women are not highly represented. Salaries are lowest in fields in which women are prevalent, such as life sciences and social sciences. Within each of these fields, the salaries of men and women are similar among those less than 30 years old, but differences between men's and women's salaries increase with increasing age. Such factors as number of years in the labor force, primary work activity, supervisory status, and number of people supervised also influence salaries and may account for some of the gap. The following section examines the influences on doctoral salaries, many of which also influence the salaries of those with bachelor's and master's degrees. The Doctoral Gender Salary Gap
In 1993, among employed science and engineering doctorate-holders  who worked full time,  the average salary for women was $50,200 compared with $63,600 for men.  (See text table 5-2.) The observed gender salary gap of $13,300 is quite substantial and corresponds to women's making only 79 percent of what men make. As has been documented in this report, however, many differences between men and women in the doctoral labor force help explain this salary gap,  e.g., women are, on the average, younger than men and have more frequently majored in fields such as the social sciences that have relatively low pay.
To determine how much of the $13,300 doctoral gender salary gap could be "explained" by differences between men and women on characteristics expected to affect their salaries, a statistical analysis was performed. This analysis permitted estimation of how large the salary gap would be if men and women in the doctoral labor force were similar on a large number of variables-the year the doctorate was received, science and engineering degree field, other work-related employee characteristics, employer characteristics, type of work performed, and indicators of "life choices." Together, these variables accounted for an estimated $11,900 of the observed $13,300 difference between the average salary of male science and engineering doctorate-holders and the average salary of female science and engineering doctorate-holders. The variables examined failed to explain the remaining $1,400 of the gap. This residual gap could have a number of possible causes:
- Although most of the important nondemographic factors that one would expect to affect differentially the salaries of men and women doctorate-holders were statistically controlled, it was not possible to control for all such factors.  Among the variables that would be interesting to add in the future are
- measures of productivity, such as the number of books and articles published; 
- prestige of the school or department from which the individual received his or her degree; 
- prestige of the school or department at which employed;  and
- more direct measures of the importance of salary as a factor in job selection.
- The measures of the variables examined are imperfect. Better measures of some of the variables might add to the ability to explain the gender salary gap. For example, 20 categories were used to measure degree fields. Within each of these degree fields, however, the subfields may differ from one another in terms of salary and gender representation.
- The results are also potentially influenced by other types of errors such as sampling error and nonresponse bias that are inherent in sample surveys. 
- Some or all of the "unexplained" gender salary gap may be attributable to "unequal pay for equal work." Indeed, the size of the unexplained gap may even be underestimated. For example, it is possible that chance has led to the inclusion of a disproportionately high percentage of high salaried women in the sample. Further, one can argue that some of the "explanatory" variables included in the analysis should have been excluded. For example, if one believes that the primary reason that women are less likely than men to go into certain fields is a perception that these fields are inhospitable to women, one might argue that field of degree should not be used as an "explanatory" variable when examining the salary gap between men and women.
In the remainder of this section, more detail is presented on the importance of the variables examined in contributing to the explanation of the gender salary gap.
Years Since Receipt of Doctorate
In the earlier chapters of this report, a long-term increase in the percentage of science and engineering doctoral degrees going to women was noted. Although this can be viewed as progress, it also means that women doctorate-holders are, on average, more recent doctorate recipients than are men. In 1993, the average full-time employed woman science and engineering doctorate-holder had received her doctorate approximately 10.4 years ago, compared to the average man who had received his degree approximately 15.7 years earlier. (See appendix table 5-32.) The gender difference in years since receipt of the doctorate "explains" approximately $3,200 of the observed $13,300 salary gap. (See text table 5-2.) This means that the difference in years since receipt of the doctorate accounts for almost one-quarter of the observed gender salary gap.
Field of Degree
Field of degree varies considerably between men and women. Women in the doctoral science and engineering population are disproportionately concentrated in psychology and the social sciences, whereas men are disproportionately represented in physics and engineering (see appendix table 5-32). Because science and engineering degree field is an important determinant of salary for the doctoral population, this variable may be helpful in explaining the gender salary gap. As seen in text table 5-2, it explains approximately $1,500 (11 percent) of the observed gender salary gap. 
Several variables on the 1993 Survey of Doctorate Recipients (SDR) that measure attributes of the individual's background prior to degree completion may affect salary. These variables are mother's education, father's education, and whether the individual lived in a rural area during the time he or she was growing up. None of these variables had a statistically significant impact on salary and, therefore, were not included in the final analysis. 
Other Work-Related Employee Characteristics
Individuals can, of course, enhance their job skills subsequent to receipt of the doctorate. They can engage in additional educational and training activities, obtain work experience, and participate in professional society activities. The SDR contains a considerable number of relevant measures to use in examining the impact of these variables on the gender salary gap. These include type of additional degrees (e.g., none, M.D., law degree) received since the science and engineering doctorate, whether the individual has taken additional courses since the last degree, the number of years of full-time work experience, whether the individual attended any professional society meetings or conferences within the last year, and the number of national or international professional society memberships.
Other work-related employee characteristics that are included in the SDR and that are associated with salary are age at time the doctorate was received, whether the individual has previously retired,  whether the individual has a license related to his or her occupation, whether the individual was employed in 1988, and if so, whether he or she has changed occupations since 1988. 
Text table 5-2 shows that these additional employee characteristics add considerably to an understanding of the gender salary gap. Collectively, they explain approximately $2,500 (19 percent) of the gap. Most of this explanatory power (13 of the 19 percentage points) is attributable to differences between men and women in years of full-time work experience. (See appendix table 5-32.) Also worthy of note is that age at time the doctorate was received explains approximately 5 percent of the gap, even though the difference in age between men and women at the time of degree is fairly small (33 years for women compared with 31 for men).
Women science and engineering doctorate-holders are less likely to be employed in the private sector, where salaries are relatively high-21 percent of the women in this analysis were employed in this sector compared with 33 percent of the men. (See appendix table 5-32.) We therefore expect differences in the type of employers to help explain the gender salary gap.  A second employer characteristic of relevance to salary analysis is the region of the country in which the employer was located-though the differences between men and women on region of employment are small. These two variables accounted for $1,300 (10 percent) of the doctoral gender salary gap.
Type of Work
A number of variables in the SDR permit examination of gender differences in type of work performed. These include occupation, whether the occupation is closely related to the degree received, primary and secondary work activities, whether the position is a management position, the number of employees supervised directly, the number supervised indirectly, and whether the position is a postdoctoral appointment. These variables jointly explain approximately $2,000 (15 percent) of the doctoral gender salary gap. None of the individual variables within this group was responsible for more than 4 percentage points.
The last set of variables consists of those labeled "life choices." Jobs typically entail a number of rewards in addition to salary (such as fringe benefits and prestige) and also entail costs, such as the opportunity costs associated with the time spent on the job. Employers are likely to find that they can offer relatively low salaries to fill positions with high nonsalary rewards or low nonsalary costs. Men and women may place different values on these nonsalary aspects of jobs, and this may result in salary differentials. For example, if, on the average, women place a higher value on having a "short" work week than do men (e.g., because of greater responsibilities for child care), women may be more likely to choose positions with relatively low salaries and fewer work hours per week.  Although the SDR does not directly ask individuals to rate the importance of different factors in their job selection, a number of variables on the database are relevant for an understanding of these "life choices."
Variables in the "life choices" set include family-related variables-marital status; whether spouse was working full time, part time, or not at all; and whether spouse had a position requiring at least bachelor's-level expertise in the natural sciences, computer science, or engineering. Also included in this category are reasons related to why individuals took the following actions: worked outside of the field of doctorate, changed occupation or employer between 1988 and 1993, took courses following completion of the most recent degree, and took work-related workshops or other training.
The variables in this group collectively explain $1,400 (11 percent) of the doctoral gender salary gap. Seven of the 11 percentage points were accounted for by marital status (see appendix table 5-32). Women were much less likely than men to be married (63 percent compared with 83 percent); being married had a positive effect on salary.
In sum, the salary gap is substantial between men and women with science and engineering doctorates, but approximately 90 percent of the observed $13,300 gap can be accounted for by differences between men and women on the variables examined in this analysis. The most important explanatory variable is years since receiving the doctorate, a variable that explains $3,200 of the observed salary gap. A wide variety of employee, employer, and work characteristics also contribute to the explained salary gap. The remaining $1,400 (10 percent of the observed gap) that is not accounted for by the statistical analyses examined in this chapter can be interpreted as an estimate of employer preferences for different types of employees. It is important to recognize, however, that it is, at best, a rough estimate, because statistical models are never able to capture with complete accuracy the true complexity of human behavior.
 For 1991 figures, see Women, Minorities, and Persons With Disabilities in Science and Engineering: 1994, p. 95.
 U.S. Department of Labor, Bureau of Labor Statistics, Occupational Outlook Handbook, 1994-95. May 1994, Bulletin 2450.
 Secondary science and mathematics teaching is not considered employment in science or engineering because most who are employed in this area have degrees in education, not in science or engineering. Only 29 percent of the science and mathematics secondary teachers responding to the National Survey of College Graduates had degrees in science or engineering.
 Committee on Women in Science and Engineering, National Research Council. 1994. Women Scientists and Engineers Employed in Industry: Why So Few? Washington, DC: National Academy Press.
 Preston, Anne E. 1994. Presentation on "Occupational Departure of Employees in the Natural Sciences and Engineering" cited in Committee on Women in Science and Engineering, National Research Council Committee on Women in Science and Engineering. 1994. Women Scientists and Engineers Employed in Industry: Why So Few? Washington, DC: National Academy Press.
 The unemployment rate measures the percentage of those in the labor force who are not employed but are seeking work.
 The difference in unemployment rates is statistically significant, i.e., it is larger than expected from chance fluctuations.
 Anne Preston, "A Study of Occupational Departure of Employees in the Natural Sciences and Engineering," National Research Council Committee on Women in Science and Engineering conference, Irvine, CA, January 17, 1993.
 Federal Glass Ceiling Commission, "Good for Business: Making Full Use of the Nation's Human Capital," March 1995. U.S. Department of Labor. Washington, DC. See also Committee on Women in Science and Engineering, National Research Council, Women Scientists and Engineers Employed in Industry: Why So Few? 1994. Washington, DC: National Academy Press.
 U.S. Department of Education, National Center for Education Statistics. 1996. Institutional Policies and Practices Regarding Faculty in Higher Education Institutions, 1992.
 The salary gap analysis focuses only on the doctoral salary gap. The salary gaps for those with bachelor's and master's degrees are, of course, also of interest, but time limitations and data availability did not permit such analyses for this report.
 Those sections of this chapter that analyze the salary gap exclude those who are self-employed and those who work part time, because annual salaries for part-time or self-employed work are not strictly comparable to full-time salaries. See the chapter 5 Technical Notes for information on how salary and some of the other variables were measured in this analysis.
 This analysis uses the 1993 Survey of Doctorate Recipients. It builds on an extensive literature in which the issue of the salary gaps for different populations is examined. See Blau and Ferber (1986) for an overview of literature on the gender salary gap.
 To examine the issue of salary equity, we use statistical techniques that permit a more comprehensive approach than is possible using the cross-tabulation approach used in most of this report. These techniques are discussed in the chapter 5 Technical Notes.
 See the chapter 5 Technical Notes for a discussion of how variables were selected for inclusion in the final model.
 Broder (1993) points out that this is a frequently used measure in the analysis of salary differentials in the academic labor market.
 Interestingly, Formby et al. (1993) did not find this variable significant in their analysis of the entry-level salaries of academic economists. Clark (1993), however, found significant impacts of both quality of granting institution and quality of employing institution on salary.
 Broder (1993) found an insignificant salary premium for prestige of the university in her sample of economists. Formby et al. (1993), however, found this variable to be highly significant. The type of academic institution, as measured by Carnegie code, is, in part, a measure of prestige; however, there are more refined measures available, though none that were mapped to the 1993 Survey of Doctorate Recipients at the time this analysis was performed.
 See Guide to NSF Science and Engineering Resources for an overview of the methodology employed in the 1993 Survey of Doctorate Recipients, including possible sources of error.
 For the purposes of this presentation, we have included in the broad field of degree category a set of variables that reflect the fact that the effect of years since doctorate on salary is not necessarily the same for all degree fields. These interaction effects explain 9 percent of the salary gap, i.e., equalizing women and men on these interaction variables would lead to an increase in the salary gap. The main effect of field of degree is a 20 percent decrease in the gap. (See appendix table 5-32.)
 This methodology is discussed in the chapter 5 Technical Notes.
 "Retired" individuals are included in the present analysis only if they were working full time in April 1993.
 See the chapter 5 Technical Notes for information on variables excluded from the analysis because there was not a statistically significant relationship.
 See the chapter 5 Technical Notes for a discussion of how type of employer is measured.
 See Barbezat (1992) for an analysis of the relationship between gender and choices among PhD graduate students in economics who were seeking employment in 1988-1989. Most important for the present analysis was her finding that men rated the importance of salary and fringe benefits of prospective employers significantly more highly than did women.