Gender Biases in Job Referrals

Abstract

Job referrals through informal networks are an essential channel for disseminating information about the qualifications of job candidates. As such, they play a crucial role in determining the outcomes of hiring and promotion decisions. In this paper, we study gender biases in the referral process. We investigate this question through an online experiment in which university students are asked to nominate their highest-scoring classmates in either a math or a verbal task. Using administrative data, we reconstruct the students’ co-enrolment network. This allows us to identify who is chosen as well as everyone else who was not. In other words, we can measure the quality of the referrals and the characteristics of candidates who are better but not chosen. We find that participants are more likely to refer men than equally qualified women in the math task but not in the verbal task. This difference is partly explained by gender differences in network structure, i.e., who is linked with whom. However, equally important are gender biases in the referral of known contacts. Thus, debiasing the referral process could substantially increase the share of women being referred.