Recent studies have found that in the United States, women earn between 71-74 cents for every dollar earned by men.[2] However, such findings often do not take into account a variety of factors that alone or together may explain much of this gap between men’s and women’s salaries.

This issue brief examines the gender salary gap in engineering, an occupation in which women held 10 percent of the jobs in 1995.[3] Using multivariate regression analysis, the authors explored various potential explanations for the salary gap in this occupation. The study showed that the salary gap is primarily explained by the fact that female engineers, on average, have fewer years of experience since their first baccalaureate degree than males; salaries of female and male engineers with similar years of experience are virtually the same.

Using 1995 survey data from the National Science Foundation,[4] the study focused on U.S. residents who were employed full-time as engineers; these included approximately 1.5 million college graduates of all ages in 16 engineering occupations.[5] Within this population, the median salary for women was 13 percent less than the median salary for men; i.e., women in engineering earned 87 cents for every dollar earned by men, higher than that for the general population. This is consistent with other results that show that the earnings gap lessens significantly when women are compared with men with similar educational backgrounds and occupations.[6]

The analysis first explored whether men and women differed on a number of measures such as educational field (engineering vs. non-engineering), employment sector, geographic region, engineering specialty, highest degree attained, and years of experience. Regression analysis was then used to determine whether these factors, alone or together, could help explain the gender salary gap among engineers. This technique provides an estimate of the size of the salary gap if men and women were identical on these measures.

Occupational Characteristics of Males and Females
Men and women differed on these occupational characteristics:

• Engineering education. Although most engineers had earned engineering degrees, a smaller percentage of female engineers had earned engineering degrees than male engineers (65 percent of women vs. 78 percent of men).
  
  
• Employment sector. There was a slight difference between men and women with respect to the sectors in which they worked. Women were better represented in local, state and Federal Government (17 percent of women vs. 13 percent of men) and less well-represented in private industry[7] (77 percent of women vs. 81 percent of men).
  
  
• Region. Men and women were distributed differently in terms of residence, e.g., a higher fraction of female engineers lived in the Pacific Coast area (24 percent of women vs. 20 percent of men).
  
  
• Engineering specialty. The one specialty that showed greater female concentration was computer software engineering (26 percent of women vs. 13 percent of men). Most other fields showed greater male representation, e.g., electrical and electronics engineering (12 percent of women vs. 21 percent of men) and mechanical engineering (10 percent of women vs. 17 percent of men).
  
  
• Highest degree attained. The only notable difference was at the doctoral level with a higher percentage of men (5 percent) having doctorates than women (3 percent).
  
  
• Years of experience. Using the time elapsed from the year of an individual’s first bachelor’s degree award as a proxy for years of experience, figure 1 illustrates the different distributions of years of experience of male and female engineers. Higher percentages of female engineers were found among those with 12 or less years of experience. On average, women in engineering occupations had five fewer years experience than men. Figure 2 shows the relationship between this measure of experience and median salary. As expected, salary levels increase with years of experience. Further, the rate at which salary increases with experience is the same for men and women.

Findings
The analysis shows that when controlling for years of experience in the regression, the estimated difference in salaries between men and women fell from 13 percent to 3 percent, bringing women’s median earnings to 97 cents to the men’s dollar. When the other variables listed above are added to the regression, the estimated difference is lowered only another 1 percentage point.

The remaining 2 percent difference cannot be explained with the available variables. Although statistically significant, the remaining difference is small compared to sources of errors such as the tendency to round salaries to the nearest $1,000, and the possible effect of factors not covered by the survey.

Limitations of the Analysis
The proxy for experience in this study would overstate the experience of those who may have temporarily left the workforce since graduating from college. Thus, if women’s participation in the engineering workforce is more intermittent than men’s, this analysis would underestimate that difference.[8]

Further research in this area needs to consider other circumstantial factors that could enhance explanation of the gender salary gap in engineering. These factors include, but are not limited to, degree combinations,[9] worker productivity, quality of the school or department from which the individual received degree(s), quality of the employer, and lifestyle or family-related choices.[10]

We could not fully explore why there are fewer older women in the engineering workforce. One factor is a lower proportion of female engineering graduates 30 years ago than in recent years. However, the analysis also showed that women with engineering degrees were working in occupations other than engineering more often than men. Further analysis is needed to explore these issues.

Copies of reports related to the topic of this Issue Brief can be obtained from:

Lawrence Burton
Division of Science Resources Studies
National Science Foundation
4201 Wilson Boulevard, Suite 965
Arlington, VA 22230
703-306-1774 ext. 6913
E-mail: lburton@nsf.gov

SRS data are available through the World Wide Web (http://www.nsf.gov/statistics/). For more information about obtaining reports, contact pubs@nsf.gov. or call (301) 947-2722. For NSF's Telephonic Device for the Deaf, dial (703) 306-0090. In your request, include the NSF publication number and title, your name, and a complete mailing address.