Anticipated Discrimination and Gender Stereotypes in Job Applications - Experiment 3

We will run two treatments which vary the gender of the participant for which the current subject is matched: male or female.

2026-09-02

2026-09-24

Selection decisions of participants as managers and recruiters

We ask participants 4 belief questions. Each is incentivised and one is
chosen at random to be payoff relevant.
1. The 50th percentile performance of men in the final round.
2. The 50th percentile performance of women in the final round.
3. The perceived 50th percentile performance of men in the final round by the
male/female managers (depending on treatment).
4. The perceived 50th percentile performance of women in the final round by the
male/female managers (depending on treatment)

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

In a previous experiment, we hired participants to act as workers. These workers completed a maths-memory task over 3 rounds. In each round, the difficulty of the task varied together with the incentive scheme. Round 1 served as a practise round in which participants played 1 minute in easy mode and 1 minute in hard mode, earning a piece rate amount per point scored. In round 2, workers were randomly assigned to have either a mostly easy round or a mostly hard round and completed their assigned tasks for 2 minutes. This round was paid as a competition: the top scorer in a randomly formed group of 4 earned a high piece rate per point scored while the others in the group scored nothing. In round 3, all workers completed the task on hard mode for 2 minutes. The remuneration scheme was a choice: workers could choose the round 1 piece rate payment or receive a high piece rate per point scored if and only if they would be selected by a recruiter. Two other participants, a manager}and a recruiter chose one worker per group observing their round 2 score and a gender revealing avatar. The manager was incentivised to select the worker that they believed would score the highest in the group in round 3. The recruiter was incentivised to select the worker that they believed the manager would select. All participants knew that the manager was a male.

In the current experiment, participants act as hiring managers and recruiters. Each participant evaluates 10 randomly matched worker groups of four in the role of manager, with the required worker data being taken from a previous experiment of ours. In each worker group, the participant observes the round 2 performance scores of the four workers. The manager chooses one worker in each group and receives a bonus of £1 if no worker in that group achieves a higher round 3 score than their chosen one.

On top of these 10 manager decisions, the participant acts as a recruiter. They evaluate the same 10 groups of four displayed in a random ordering with the four workers arranged differently. The recruiter chooses one worker in each group and receives a bonus of £1 if a randomly selected male/female manager making the same decision chose the same worker as them.

The participant plays in these two roles in random order; about half of the participants will first act as managers and then as recruiters, while the other half does the opposite order. This allows us to do both a within- and a between-participant analysis of manager and recruiter stereotypes in hiring decisions. Aside from varying the roles in which they are acting, participants are randomised to a treatment determining the gender - male or female - of the manager for whom their decision is incentivised as a recruiter.

Computer randomisation

Participant

No

580 participants

580 participants make 20 decisions indicating 11600 choices.

145 by gender and treatment