Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper studies the relationship between pre-college skills and gender differences in STEM majors. I use longitudinal data to estimate a generalised Roy model of initial major choices and subsequent graduation outcomes. I recover students’ latent math ability, non-cognitive skills and math self-efficacy. High–math-ability women have lower math self-efficacy than men. Mathematical ability and self-efficacy shape the likelihood of STEM enrolment. A lack of math self-efficacy drives women’s drop out from STEM majors. I find large returns to STEM enrolment for high–math-ability women. Well-focused math self-efficacy interventions could improve women’s STEM graduation rates and labour market outcomes.