Beliefs about gender bias and hiring decisions
How to cite
Thunstrom, Linda; Mollerstrom, Johanna; Lindley, Sarah; Ashworth, Madison. 2024. "Beliefs about gender bias and hiring decisions." AEA RCT Registry. AEARCTR-0013679. https://doi.org/10.1257/rct.13679-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
People may expect gender biases in contexts where there is none, and such biased expectations may be consequential. We design an experiment to test whether inaccurate beliefs about the existence of gender biases may cause employers to favor men employees. Our experiment has three groups of agents: experts, experts’ employers, and customers. The employers’ task is to hire either a man or a woman expert with identical credentials. The employers’ payoff depends on the customers’ willingness to pay (WTP) for the expert’s service (advice). Hence, the employers’ expected payoff, and therefore the payoff from their hiring decision, depends on their expectations of customers’ WTP for advice from a woman versus a man expert. Customers may buy expert advice in an incentivized Becker-DeGroot-Marschak (BDM) auction, to help with their task in the experiment. The customers’ task is to predict the inflation rate in the U.S. economy, and the customer whose prediction is the most accurate gains a relatively large monetary reward. The payoffs of the employers in the experiment are designed to eliminate the existence of altruistic motives for hiring experts, and we control for other potential mechanisms causing employers to favor any particular gender, beyond beliefs about gender biases amongst customers.
Intervention
Our study has three groups of agents: economic Experts, experts’ Employers, and Customers. We are interested in the behavior and beliefs of Customers and Employers.
Customers` participate in an RCT, and are tasked to make a prediction of the inflation rate for 2024. The Customer who makes the most accurate prediction (as determined by the comparison of the prediction to the official actual inflation rate in May -- a number made official by US govt agencies in mid June 2024) wins a bonus of $200. Customers therefore have an incentive to make the best possible prediction.
To help them successfully predict the inflation rate, Customers can buy advice from economic Experts, consisting of the Expert's prediction of the inflation rate in May 2024. Each Customer is randomly paired with either a man or a woman Expert, from whom they can buy advice. This advice consists of a prediction of the inflation rate in May made by the economic Expert. Economic Experts are identical in their qualifications, but differ in gender. Customers' WTP for advice is elicited in a Becker-DeGroot-Marschak (BDM) auction. This design allows us to examine whether Customers do exhibit a gender bias, i.e., differ in their WTP for advice from a man vs a women Expert, even though the Experts have identical qualifications.
Experts only role in our study is to make a prediction of the inflation rate for May 2024 that gets provided to the Customer they are paired with, should that Customer win the BDM auction.
Employers’ task is to hire an Expert. They can choose between hiring either a woman or a man Expert, with identical credentials. The Employer has an incentive to hire the Expert that they believe will generate the highest WTP from Customers -- the Employers’ payoff depends on Customers’ willingness to pay (WTP) for the Expert’s service (advice). We elicit Employers' incentivized beliefs about Customers' WTP for advice from a man and a woman Expert. We also elicit Employers' own beliefs about the value of Expert advice from a man vs a woman Expert. This design allows us to examine if Employers believe that Customers are biased against women Experts, and therefore choose to hire a male Expert, even if they themselves do not exhibit any gender bias.
We will also separately collect data from Employers where we correct for their biased beliefs, i.e., inform them about the Customers' actual WTP for advice from a man vs a woman Expert. This enables us to examine whether correcting for biased beliefs of any gender bias amongst their Customers changes their willingness to hire male and female Experts.
Experimental design
Our study has three groups of agents: economic Experts, experts’ Employers, and Customers. We are interested in the behavior and beliefs of Customers and Employers. Our participants get randomized into the role of Customer or Employer.
Customers` participate in an RCT, and are tasked to make a prediction of the inflation rate for 2024. The Customer who makes the most accurate prediction (as determined by the comparison of the prediction to the official actual inflation rate in May -- a number made official by US govt agencies in mid June 2024) wins a bonus of $200. Customers therefore have an incentive to make the best possible prediction.
To help them successfully predict the inflation rate, Customers can buy advice from economic Experts, consisting of the Expert's prediction of the inflation rate in May 2024. Each Customer is randomly paired with either a man or a woman Expert, from whom they can buy advice. This advice consists of a prediction of the inflation rate in May made by the economic Expert. Economic Experts are identical in their qualifications, but differ in gender. Customers' WTP for advice is elicited in a Becker-DeGroot-Marschak (BDM) auction. This design allows us to examine whether Customers do exhibit a gender bias, i.e., differ in their WTP for advice from a man vs a women Expert, even though the Experts have identical qualifications.
Experts only role in our study is to make a prediction of the inflation rate for May 2024 that gets provided to the Customer they are paired with, should that Customer win the BDM auction.
Employers’ task is to hire an Expert. They can choose between hiring either a woman or a man Expert, with identical credentials. The Employer has an incentive to hire the Expert that they believe will generate the highest WTP from Customers -- the Employers’ payoff depends on Customers’ willingness to pay (WTP) for the Expert’s service (advice). We elicit Employers' incentivized beliefs about Customers' WTP for advice from a man and a woman Expert. We also elicit Employers' own beliefs about the value of Expert advice from a man vs a woman Expert. This design allows us to examine if Employers believe that Customers are biased against women Experts, and therefore choose to hire a male Expert, even if they themselves do not exhibit any gender bias.
Design details
Cited by
No citation data available for this registration yet.