The match between ‘She’ and ‘Her’: Performance gains from gender match in higher education
Abstract
I use a novel and confidential administrative dataset from a highly rated economics program in Bangladesh to show that female students taught by female instructors gain approximately 9 percent of a standard deviation in grade performance relative to their male counterparts. In addition, I find suggestive evidence that female students benefit more than male students from cumulative exposure to female instructors in longer-term outcomes, such as degree completion time and the likelihood of enrolling in an economics master’s program. The quasi-random allocation of students to mandatory courses, where students cannot select either courses or instructors, alleviates self-selection concerns. I rule out the explanation that the gain is driven by gender preference in teachers’ assessment and find that the gain appears to be driven by instructor gender itself rather than by teacher attributes correlated with gender, as the effect remains large and statistically significant even after controlling for observable teacher characteristics.