Anticipated Discrimination and Gender Stereotypes in Job Applications - Experiment 2

How to cite

O'Leary, Ethan; Caputo, Roberto; Hilweg-Waldeck, Michael; Zednik, Anita. 2026. "Anticipated Discrimination and Gender Stereotypes in Job Applications - Experiment 2." AEA RCT Registry. AEARCTR-0019233. https://doi.org/10.1257/rct.19233-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Ethan O'Leary* , Roberto Caputo , Michael Hilweg-Waldeck , Anita Zednik

* Primary investigator

Registry ID
AEARCTR-0019233
DOI
https://doi.org/10.1257/rct.19233-1.0
First registered
Last updated
Status
In development
Trial dates
2026-08-11 – 2026-08-25
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/19233

Abstract

We study to what extent including evaluator subjectivity in tournament decisions affects the documented gender gap in competition. We conduct a large-scale, pre-registered online experiment ($n=7310$) in which workers choose between a non-competitive piece rate and a competitive payment scheme whose payout depends on being selected by an incentivised human evaluator. By randomly varying whether the evaluator can observe the worker's gender across a male-stereotyped and a non-stereotyped task, we exogenously induce the scope for anticipated discrimination. In the male-stereotyped task, gender revelation breaks the link between women's objective performance beliefs and their beliefs about being selected: conditional on the subjective probability of having the best signal, moderately confident women report significantly lower selection expectations when their gender is known. This distortion reduces female entry rates, with the strongest effects among non-risk averse participants: those who made up the majority of the competitor pool when their gender was concealed. The revelation of gender in the male-stereotyped domain particularly effects low-ability women, destabilising the negative association between ability and competition among females. We validate that this effect is driven by the anticipation of gender discrimination and not by a preference for opting out of settings in which gender is made salient: in a non-stereotyped task we find no change in beliefs nor in competition behaviour when their gender is revealed to the evaluator. Our findings suggest that the extent to which the gender competition gap distorts female labour supply may be underestimated in the face of the potential for evaluator bias, even before it is enacted.

Experimental design

The experiment involves selection of workers from a menu according to who is most likely to perform in the future.

Design details

Randomization method
Computer randomisation
Randomization unit
Participant
Clusters
900 (450 per treatment)
Observations
18900 - 21 menus per participant, each menu contains 4 discrete choices: 21 x 4 x 900 = 18900
Arms
450 participants per treatment so 9450 observations per treatment
Primary outcomes
Selection decisions of managers
Secondary outcomes
We collect data on beliefs in the final survey part of the experiment. First we measure participants’ beliefs about the distribution of round 3 scores for men and women separately through the following three distributional moments: (i) median, (ii) the 80th percentile, and (iii) the 20th percentile. The elicitation procedure is as follows. We state that we randomly draw 10 participants of the gender of interest and order them by their round 3 score. We then ask managers to state a number which is (i) no less than than score of the 5th placed worker and no more than the 6th placed worker, (ii) no less than than score of the 8th placed worker and no more than the 9th placed worker, and (iii) no less than than score of the 2nd placed worker and no more than the 3rd placed worker. Next, we measure their belief on how informative and biased the round-2-signal of workers is for each gender. To do so, we inform participants that we randomly draw 10 participants of the gender of interest who each had a round-2 signal of 16 (the mode) and order them by their round 3 score. Among this line-up, we repeat the previous 3 questions where (i) measures the perceived bias and (ii)-(iii) is an indicator of the perceived precision of the round-2 score as a signal of the round 3 score.

Cited by

No citation data available for this registration yet.

Keywords