Gender differences in student evaluations of teaching: identification and consequences

Student Evaluations of Teaching (SETs) have been suggested as one possible cause for low representation of women among academic economists. While econometric analyses using control variables certainly report that SETs can be influenced by the gender of both teacher and student, such studies may still be biased if there is non-random allocation of teachers to teaching. Even if causal estimates of gender effects are unbiased, the inference that SETs contribute to gender discrimination is hazardous, since hiring or promotion committees would not have access to these controls when evaluating SETs. We use data from an Italian university to quantify the effect of controls on gender effects and conclude that there is insufficient evidence to blame SETs for a gender imbalance in Economics.