Masked Evaluations: The Role of Gender Homophily

Abstract

Removing demographic information from job applications has been a proposed solution for reducing gender discrimination and increasing diversity in the hiring process. In this paper, I use a unique data set of job application questions from a hiring platform to explore how gender might still impact masked evaluations. I find that while there is no main gender effect in the scoring of applications, there appear to be scoring patterns consistent with gender homophily. Having applicant-evaluator pairs being of the same gender is positively correlated with higher ratings, particularly for female applicants, and neither the applicant’s language nor other types of similarity appear to fully explain this association.