Gender Bias in Beliefs about Performance

Last registered on April 26, 2024


Trial Information

General Information

Gender Bias in Beliefs about Performance
Initial registration date
April 20, 2024

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
April 26, 2024, 11:29 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.


Primary Investigator

The Ohio State University

Other Primary Investigator(s)

PI Affiliation
Loyola Marymount University

Additional Trial Information

In development
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
When mean beliefs about men and women are elicited, often there is very little difference. But in line with previous literature on perceived variance of men vs. women, we aim to investigate whether the extremes of belief distributions over performance inform hiring decisions. Our laboratory experiment is designed to first test whether men and women are hired disproportionately for jobs with different expectations of performance. Additionally, we elicit beliefs about the variance of men and women's performance in order to determine whether the shape of belief distributions gives insight into why we might see a gender gap in hiring.
External Link(s)

Registration Citation

Leo, Greg and Samantha Stelnicki. 2024. "Gender Bias in Beliefs about Performance ." AEA RCT Registry. April 26.
Experimental Details


Participants in a laboratory experiment will decide whether to hire a randomly chosen man or woman worker for three different jobs that are based on worker performance on a math task. Each participant hires a worker for all three jobs -- the order of which is randomized at the participant level. We are interested in the difference of number of men and women hired across each of these tasks, with our intervention being within-subject. Additionally, separate individuals will be asked their beliefs about the performance of a randomly selected worker on the math task. These individuals will be randomly placed into one of two treatments. In the men treatment, participants are asked about their beliefs of a randomly chosen male worker on the math task. In the women treatment, participants are asked about their beliefs of a randomly chosen female worker on the math task.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
We have three main outcomes of interest. The first is differences in the percentage of men and women hired for each job in the Hiring Task. The second is across each job, comparisons of the number of men hired and comparisons in the number of women hired. Both of these first two outcomes will be investigates using Chi-Squared tests for differences in proportions. The third outcome of interest is the difference in beliefs about men and women's performance. This is the difference in average number of men and women reported in each belief question. In order to analyze this difference we will use a t-test for differences in means.
Primary Outcomes (explanation)
We have no constructed outcomes. All outcomes are comparisons of unmodified aggregated data.

Secondary Outcomes

Secondary Outcomes (end points)
Our secondary outcomes will include comparisons of actual performance with beliefs about performance of workers, correlation between beliefs and hiring decisions at a population-level, and whether the gender of the hiring manager leads to differences in gender bias in hiring task.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The laboratory experiment consists of three stages: Worker Stage, Beliefs Stage, and Hiring Stage. In the Worker Stage, 20 workers will be recruited to take a math task that will be used as the basis for Beliefs Stage and Hiring Stage decisions. In the Beliefs Stage, beliefs about men's and women's performance on the math task will be elicited. In the Hiring Stage, managers will hire workers for three different jobs based on math task performance. Further details about each stage are explained in the following sections.

Worker Stage:

20 participants, 10 men and 10 women, will perform a math task. Before the math task, we ask for demographic information, including age, gender, major, and year. The math task, itself, is the addition of five two-digit numbers. Participants have 5 minutes to answer as many math questions as possible. Payment for the math task is piece-rate, where each correctly answered question is worth $1.00. Participants will be told prior to participating that their demographic information and performance will be shown to other people in a later study.

Beliefs Stage:

In the second part of the experiment, participants will be recruited to tell us their beliefs about performance of others on the math task. Half the participants will tell us their beliefs about the distribution of men's performance on the math task and half the participants will tell us their beliefs about the distribution of women's performance on the math task. Participants will be randomly assigned to either the men questions or women questions.

We will elicit beliefs using a Multiple Price List about the probability that one randomly chosen man/woman in the worker stage is in the bottom 25th, bottom 50th, and top 25th quantiles of performance in the population of workers. This question will be: What is the probability that a randomly chosen (fe)male worker performs in the top 5 of total workers' performance? We will ask three questions where we replace "top 5" with "bottom 5" and "bottom 10".

Hiring Stage:

In the hiring stage, managers to hire from the previous pool of workers. There are three different "jobs" for which managers will hire. For the first job, managers will get paid if the worker they hire "passed" the math task. Passing the math task means that the worker did not score in the bottom 25th percentile of performance on the math task. For the second job, managers will get paid if the worker is excellent on the math task. Excellence means that the worker scored in the top 25th percentile of performance on the math task. The last job is a piece-rate payment for the manager based on the performance of the worker. This means the manager will receive $1.00 for each question the worker answered correctly. For each of these three questions, the manager will choose between one randomly selected man and one randomly selected woman from the worker population.
Experimental Design Details
Not available
Randomization Method
Randomization happens electronically through a random number generator integrated in the experimental programming.
Randomization Unit
Randomization happens at the individual level.
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
No planned clusters.
Sample size: planned number of observations
20 observations of performance on math task, 240 observations of beliefs, and 300 observations of hiring decisions.
Sample size (or number of clusters) by treatment arms
20 workers, 80 adults in Beliefs Task, 100 adults in Hiring Task
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
For our three main planned analyses, we calculate minimum detectable effect sizes with a power of 80% and significance level of 5%. For the chi-squared test of difference in proportions, our minimum detectable effect size is 20%. For our t test of differences in mean beliefs, the minimum detectable effect size is 0.8.

Institutional Review Boards (IRBs)

IRB Name
The Ohio State University Institutional Review Board
IRB Approval Date
IRB Approval Number
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information