Association for Public Policy Analysis & Management

Although Chilean women have higher rates of graduation and educational persistence, they account for only 23% of undergraduate students in technology and basic sciences. While in college, there is evidence that they perform worse than their male counterparts and that this might be related to the gender composition of their peers.

Using information on 76,124 students from one post-secondary vocational institution in Chile, I estimate a gender peer effect linear model that links dropout to percentage of female peers in STEM and Non-STEM majors. Using major/branch fixed effects, the model exploits idiosyncratic within major/branch variation in gender peer composition to estimate the effect.

The results suggest that an increase in the percentage of female peers, so STEM majors have a 50/50 gender peer composition (increase of 31.6 percentage points) is associated with a reduction of 20% (3.95 percentage points) in female students’ dropout rate. This result supports the hypothesis that in STEM majors, female students’ educational outcomes are positively related to having more female peers in their cohorts. For males, this relationship is not significant.

Some may say that the problem is not some particular gender dynamic in STEM majors, but of male concentrated majors. Nevertheless, when the effect is calculated for STEM majors that are not male concentrated, the result holds: an increase in the percentage of female peers, so STEM majors have a 50/50 gender peer composition (increase of 31.6 percentage points) is associated with a reduction of 28% (5.49 percentage points) in female students’ dropout rate.