Women and the Leaky STEM Pipeline

Summary of the research article: Speer, J. D. (2023). Bye bye Ms. American Sci: Women and the leaky STEM pipeline. Economics of Education Review, 93, 102371.

Topics: STEM, inégalité de genre, college major

Summary by Leïla Costil, Research assistant

Women remain underrepresented in the science, technology, engineering, and mathematics (STEM) fields. Men make up about two-thirds of STEM college graduates and hold an even larger share of STEM jobs. Because STEM fields typically pay well, women’s under-representation contributes to the overall gender pay gap. By tracing the experience of women along the STEM “pipeline” from high-school to the labour market in the United States, this paper asks: what stages contribute the most to the gender gap? Read the original article.

Summary

Along the path individuals undertake to work in STEM-related jobs, women often drop out at different steps, altogether explaining the gender gap in STEM careers. By far, the two most important stages are the first major choice - where students decide on what topics to specialise in, which alone account for 56% of the gender gap - and the transition from college to career, explaining around 44% of the gap. To fully understand the “leaky pipeline”, one must focus on both the college experience and the labour market.

Data

To trace the STEM pipeline from high school to the labour market, researchers need information on school experiences, grades, test scores, college major choices, and job outcomes. In this paper, the author uses pre-college and college information from the National Longitudinal Survey of Youth’s 1997 cohort (NLSY), which is a panel data set of about 9,000 respondents born between 1980 and 1984. The respondents were first interviewed in 1997 and have been followed through the present. For the labour market stages, the author relies on the American Community Survey, combining data from 2009 to 2017 (the years which contain college majors), obtaining data on 1.2 million STEM college graduates.  

Methodology

To decompose the leaky pipeline, the author faces multiple difficulties: first, because he used two different data sources, some reweighting is needed to fully compare them. Second, the younger and older cohorts in the ACS dataset are different, so another reweighting is used. Using this, to decompose the impact of the different steps on the gender gap, the author estimates the overall evolution of the gender gap from the first stage in high school (51.45% male) to STEM occupation at age 45 (70.74% male), and computes the percentage change between each individual stage as compared to the overall change.

Findings

The author divides the STEM pipeline in six important stages: pre-college “STEM-readiness”, college attendance, initial major choice, persisting to graduation with a STEM degree, early career occupation outcomes, and career progression from early-career to mid-career jobs. 

In the first stage, the author studies how prepared students are for a STEM major when they reach the age to choose such a path. Based on the selected criteria of readiness, men are slightly more STEM-ready than women. Men are also more likely to be in the upper tail of STEM-readiness, which could be important if these are the students who not only choose STEM but are most successful in it. The gap in predicted STEM probability (0.5 percentage points), while significant, is actually a small fraction of the actual gap in STEM majoring. When studying college attendance, it must be noted that women have a significant advantage over men in college attendance and graduation. To see how this affects the pipeline to STEM jobs and the gender gap in outcomes, the author looks at how attendance patterns differ by STEM readiness. While women have a significant advantage in college attendance, the effect of this advantage on the gender gap in STEM is unclear. Much of the attendance advantage is driven by the least STEM-ready students, meaning that among those who do attend, a higher share of men are STEM-ready. 

Then, once a student decides to attend college, they have to decide on the major to pursue. This choice happens usually in the second year of university (“sophomore year”), and is not binding, so that students can still change majors afterwards. The author finds that at every level of STEM readiness, men are more likely to go into STEM majors. The gaps are especially large among the more STEM-ready students. There are clearly large differences in major choice for men and women, even conditioning on STEM readiness. 

Moreover, there are substantial gender differences in major choice even within STEM. Women tend to choose biology most often, while engineering and computer science are more common for men. When they do not choose to major in STEM fields, STEM-ready women are often found in business majors, as well as psychology and fine arts. 

Conditional on starting a STEM major, women are far less likely to graduate in STEM. This is driven by large differences in persistence among the least STEM-ready students. To summarise, men with minimal STEM preparation more often either graduate in STEM or do not graduate college at all, while women are much more likely to switch to another major and graduate. 

The researcher studies if there are significant differences in occupational choices among STEM graduates, and where those students are going if not to STEM jobs. Using the narrow definition of STEM, men with STEM degrees are about twice as likely as women to be in STEM occupations (41% vs. 20%). Women with STEM degrees are much more likely to be in medical jobs or out of the labour force than men graduating from STEM fields. 

Finally, in terms of career progression, both men and women are slightly less likely to be in STEM occupations at age 45 than at age 30. During the career stage, the gender gap shrinks from 14.4 percentage points (47.6%-33.2% at age 30) to 12.7 percentage points (44.1%- 31.4% at age 45).

Overall, when decomposing the STEM pipeline, two stages stand out: at initial major choice, the male share is 61.2%, compared with the 50.2% predicted by attendance and readiness. This stage alone accounts for 57% of the total gender gap. The other stage that stands out is the age-30 occupation, representing the transition from college to career. Among STEM graduates, men are much more likely to choose a STEM job. This stage accounts for 44% of the total gender gap. STEM-readiness accounts for only 8% of the gender gap in STEM, and persistence from initial STEM major to graduation in STEM matters a bit more (16%). Women’s higher rates of college attendance mean that this stage actually narrows the gender gap in STEM (and accounts for -14%).