Who Is Biased? Heterogeneity in Gender Bias in Evaluations

Professor review websites are widely used by students but may transmit biased information. This paper examines who generates gender bias in teaching evaluations. Using scraped instructor data from RateMyProfessor.com, I document descriptive evidence of a gender gap. I then conduct a randomized experiment in which students evaluate identical teaching content delivered with male or female voice-overs. The results show a causal gender penalty of about 0.11 points on a 1–5 scale. This penalty is concentrated among students who actively post online reviews, suggesting that bias is driven by a subset of reviewers rather than the broader student population.