A Lightweight Intervention to Decrease Gender Bias in Student Evaluations of Teaching

Abstract

Women are underrepresented as instructors in engineering, computing, and technology classes. One factor that disadvantages women in the classroom are student evaluations of teaching (SETs), as research finds they contain significant gender bias. This may contribute to the dearth of women in computing education, as SETs are used in decisions about contract renewals, hiring, tenure, and promotion. The double-bind is one cause of gender bias in SETs, meaning that it is more difficult for women than for men in leadership positions (such as being a professor) to be perceived as both competent and likable. We examine a lightweight intervention's impact on gender bias caused by the double-bind. Specifically, we conducted a field experiment in which the woman professor of a CS1 class for non-majors gave students in the intervention condition additional, positive exam feedback via email. We hypothesized this would increase students' perceptions of the professor's likability, which would then increase her SETs. We find that the intervention increased top-performing students' ratings of the professors' likability. We also find that the professor received significantly higher SETs the semester she sent the intervention emails. While women should not have to alter their behavior to accommodate students' gender biases, this intervention may be a useful survival strategy for women impacted by gender bias in SETs.