Pride and Prejudice? Structural Evidence of Social Pressure from a Field Experiment with Committees

How to cite

Schwantje, Tom; Quinn, Simon; Abebe, Girum; Gole, Tom; Brooke, Siân. 2022. "Pride and Prejudice? Structural Evidence of Social Pressure from a Field Experiment with Committees." AEA RCT Registry. AEARCTR-0009893. https://doi.org/10.1257/rct.9893-1.0

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

Tom Schwantje* , Simon Quinn , Girum Abebe , Tom Gole , Siân Brooke

* Primary investigator

Registry ID
AEARCTR-0009893
DOI
https://doi.org/10.1257/rct.9893-1.0
First registered
Last updated
Status
In development
Trial dates
2022-08-10 – 2022-08-30
Countries
Private
JEL codes
D, O
Registry record
http://www.socialscienceregistry.org/trials/9893

Abstract

This pre-analysis plan details our experimental design and pre-analysis plan for an experiment on the role of institutional design on the relative success of male and female candidates in a business plan competition. We run an experiment in which judges are randomly assigned to both a committee treatment (as opposed to judging individually) and to a prompt detailing the importance of equal opportunity for male and female entrepreneurs. We pre-specify both our main reduced form analysis and a structural model yielding as main hypothesis that when members of a committee face social pressure to agree with each other, they overweight public information, generating status quo bias. We also pre-specify a more flexible generalisation of the model to test alternate hypotheses.

Intervention

In this experiment, we test the hypothesis that when members of a committee face social pressure to agree with each other, they overweight public information, generating status quo bias. We do so by implementing a business plan competition in which judges are randomly assigned to either vote independently or as part of a committee, and to a treatment prompt alerting judges to the unique constraints female entrepreneurs face and the importance of equal opportunity. We assess whether these treatments, both independently and jointly, affect the probability a judge prefers a female over male candidate, and the probability of a unanimous decision.

Experimental design

We implement a fully randomised design. Please see the attached pre-analysis plan for a detailed overview of the experimental design.

Design details

Randomization method
Randomization done in office by a computer, after which judges are assigned an ID within the randomization based on where they sit on the day.
Randomization unit
Randomisation is done at the judge-level; we have 12 observations (assessments) per judge. Standard errors are clustered at the judge-level.
Clusters
240 Judges
Observations
240*12 = 2880 observations. Twelve observations for each of 240 judges.
Arms
60 judges per treatment arm including 45 male and 15 female judges. Each does 12 assessments for 720 assessments per treatment arm.
Primary outcomes
Our primary outcome is whether a judge votes for a female candidate over a male candidate
Secondary outcomes
Our secondary outcomes are whether or not a committee's decision was unanimous, and whether the unanimous decision was for the male or female candidate in a binary comparison.

Cited by

No citation data available for this registration yet.

Keywords