Gender Bias and Instructor Availability in Student Evaluations of Teaching in Higher Education

Abstract

We run an experiment to examine the role of instructor availability in the documented gender bias of student evaluations of teaching (SETs). We find that subjects rated female instructors more harshly when unavailable, particularly if due to work conflicts. We then analyze SETs from over 3,000 courses, and compare estimates with various controls when excluding and including availability scores. When excluding availability scores, we find white women are rated better compared to men with or without other controls. Conversely, we find lower scores for black women and adjunct instructors. However, including responses on instructor availability renders the estimates for adjuncts insignificant. It also removes the significant differences for women of color and instead we get a significant difference for all women. This suggests some faculty may offset gender bias by increasing
availability.

Keywords: Gender, Evaluations, Teaching, higher education, Productivity