Association for Public Policy Analysis & Management

Despite substantial interest in reducing the gender gap in science, technology, engineering, and mathematics (STEM), much remains to be known. In this paper, we use rich data from Mexico City on student preferences for high schools to study the determinants of the STEM gender gap and to assess the potential for various policies to reduce the gap. We are among the first to describe students’ portfolio of choices, identifying fluid preferences over STEM and non-STEM options. We also contrast vocational and professional school choices. This has been less studied in the literature and is important due to the large fraction of STEM jobs that are blue-collar, particularly in developing countries.

Under the centralized school choice policy, males are 12 percentage points more likely to be assigned to a STEM school than females, with 60 percent of this gap coming from vocational schools. 34 percent of the assignment gap is attributable to differences in placement test scores, with males scoring 0.2 standard deviations higher on the test. The remaining differences are due to gendered choices. Comparing choices, males list more STEM schools than females but most males and females list both STEM and non-STEM programs in their ranked schools. For elite STEM schools, the choice gap is partially explained by differences in STEM comparative advantage but not by differences in grades or college aspirations. In contrast, non-elite STEM schools are associated with lower grades and non-college going aspirations; these explain a large fraction of the non-elite STEM choice gap. Finally, the policy simulations reveal that the gap can be reduced by 25 percent through policies such as STEM affirmative action or choice nudges, where females are induced to move their highest ranked STEM school up by one rank. These results highlight the potential for policies at the margin to have large effects.