Bias and Discrimination Against Women and Parents in Semi‐Automated Hiring Systems

ABSTRACT Today, organizations are increasingly relying on automated hiring. The mechanization of the hiring process is assumed to render it more neutral, but a growing literature shows algorithmic decisions are as likely to be biased (Dickson, 2018). In this study, we test two types of biases: (1) gender bias; and (2) parenting bias, (i.e., whether mothers and fathers with an extended gap to care for children are penalized vis‐à‐vis those with uninterrupted employment net of equivalent high‐impact qualifications). We apply a classic counterfactual study sending gender and parenthood manipulated CVs to 211 job advertisements across three occupations (men‐dominated, women‐dominated, and gender‐balanced, to mitigate confounding variables associated with gender composition) in the United States and measure penalty‐premium bias in response rates. Our results identify semi‐automated hiring bias against parents who took leave to care for children relative to those with uninterrupted employment. Importantly, we find fathers who have an extended parental leave were the most severely penalized, followed by mothers with an extended parental leave and women and men without parental leave respectively. Ultimately, we identify gender and parenting bias in algorithmic and human hiring decisions.