Academic Aptitude Signals and STEM field participation: A Regression Discontinuity Approach

Gender disparities in STEM field participation at all levels are wide and persistent. In this paper we explore whether external signals about academic aptitude can influence female participation in STEM fields. We analyze 10 years of data on aptitude tests administered by a private university in Peru taken by 3,000 high school students each year. Prior to the test, students are asked to state their (non-binding) preferences over college majors. Admission into majors is determined on the basis of cut-off scores on the exam, which has a math and a verbal component. Using a regression discontinuity design, we find that among students whose preferred major was other than engineering, making the engineering cut-off increases the likelihood of enrolling in engineering by 10-12 percentage points. These effects are driven entirely by female students, and no effect is seen for males. We also find that women with higher scores on the verbal component are less likely to make this switch, reinforcing the idea that external signals about aptitude matter for choice of college majors. These results highlight the importance of external validation in influencing career choices in a context where social norms discourage female participation in STEM fields, and have important policy implications.