Does Domain Preference Explain Gender Gaps in Self-Evaluation?

Abstract

We study whether the estimated gender gaps in confidence and self-evaluation in male-stereotyped domains are mitigated when accounting for domain preference. We run a near-replication of Exley and Kessler (2022) where participants can additionally reveal their preference for a math & science domain. All participants then complete a math & science test and provide selfevaluations. We replicate the findings of Exley and Kessler (2022) where women have lower confidence and provide harsher self-evaluations compared to equally performing men. We additionally report two novel results. First, we find that men indicate a stronger preference for a math & science domain than women do, even controlling for ability. Second, we show that controlling for this domain preference gap does not mitigate the gender gaps in confidence and self-evaluation. Our findings are consistent with an interpretation that other characteristics of the domain itself (e.g., task stereotypes) may be primary drivers of these observed gender gaps.

Keywords: Domain preference, Gender, Confidence, Self-evaluation

JEL Classification: C91, D91, J16