Experimental Design Details
Experiment 1: Avoiding the situation where identity matters altogether
Stage 1 – gaining job experience
In this stage, participants will be randomly assigned to work in two different firms in sequence to gain job experience. Their social identity within the experiment is based on the job experience they get at the first stage. They learn that there is a creative industry which comprises four firms: A, B, C, D. Participants will be randomly assigned to work at one of these firms first and then the other. The possible combinations of work experience are: A+C, A+D, B+C, or B+D. It is not possible to have combinations A+B or C+D.
Participants in each firm will work on a group task to create a logo for a client. They will choose options for the background color, font, text position, and logo icon, and provide a short explanation for their choice. Their colleagues' suggestions for the logo, including a short explanation, will also be displayed to increase salience. This task is asynchronous, non-performative, and non-incentivized, and is intended to strengthen social identities.
Stage 2 – promotion in a new firm
In this stage, participants move to a new firm and must hire a creative professional to work for them. They will consider three candidates who differ in their previous job experience. One candidate has the same job experience as the participant, meaning they both worked for the same two firms. We interpret this as Ingroup + Ingroup. Another candidate has partially matching work history with the participant, meaning they both worked for one firm but gained another job experience in different firms. For instance, the participant has experience A+C and the candidate – A+D. We interpret this as Ingroup + Outgroup. The last candidate does not have any common job experience with the participant. We interpret this as Outgroup + Outgroup.
In the control arm, participants state their willingness-to-pay (WTP) to work with each of the three candidates, making three decisions in total. This is an implementation of Becker-DeGroot-Marschak (BDM) mechanism.
In the treatment arm, participants will simultaneously state the WTP to work with each candidate and choose wages for each candidate in a dictator game setting. This means they will state their WTPs and select wages for all three candidates.
The stated WTP to hire each candidate will influence the probability of hiring them. The higher the stated WTP, the more likely the candidate will be hired. Participants will be given detailed explanations of how their stated WTP affects the probability of hiring a candidate. Based on the stated WTPs, one of the candidates will be randomly selected and hired. The wage selected by the participant for this candidate will determine the payoff.
Treatment variation
In the control arm, we aim to measure “pure identity effect”, which is the tendency of individuals to prefer others with similar job experience without considering monetary trade-offs. We will do this by measuring the stated WTP to work with each candidate.
In the treatment arm, we aim to test if financial incentives may affect how people think about their identity. On the one hand, people often favor member of their in-group over out-group members. This tendency is commonly known as “in-group favoritism”. In our setting, this could explain a stronger willingness to hire a candidate with similar work history. On the other hand, there may be pressure to offer a higher wage to such a candidate. By measuring both WTPs and wages for all candidates, we can test if participants try to avoid hiring someone with the same job experience by calibrating their WTPs and make a lower wage offer to the candidate with non-matching job experience. This will be done by having participants simultaneously choose their WTPs for all candidates and select wages in a dictator game setting.
Stage 3 – joint work in the new firm
Based on the chosen WTPs for all candidates in the previous stage, one candidate is hired. The manager’s job is to create a task for the newly hired creative professional. The manager selected visual guidelines for logo inspirations, color palette, and font types. Subsequently, the creative professional is going to develop a logo for an account following the manager’s guidelines.
Experiment 2: Strategic belief formation about others’ identity
Stage 1 – gaining job experience
Same as in the experiment 1.
Stage 2 – promotion in a new firm
In this stage, each participant moves to a new firm and needs to hire a creative professional to work for them. They will consider two different candidates. For each candidate, they will only observe one of the two firms where the candidate has worked before. The other firm of each candidate is unknown to the participant. Based on the observed job experience, a candidate has either worked at a firm where the participant has also worked (Ingroup) or a candidate does not have any shared observable experience with the participant, i.e., the candidate has worked for a different firm (Outgroup).
In the control arm, participants will be asked about their beliefs about the unknown firm of each candidate. In other words, they are asked to make a prediction, whether a candidate has worked in the same firm (Ingroup) or a different one (Outgroup). They will do this by making a monetary bet, and if their prediction is correct, they will receive an additional monetary reward. Participants make their decisions one after another on different pages.
In the treatment arm, in addition to making a bet, the participant will also simultaneously decide on the wage they would pay to each candidate. One of the decisions is randomly selected to determine the payoff. Participants make their decisions about each candidate separately, or one after another on different pages.
In the treatment, we aim to test if financial incentives can distort the belief formation. In the baseline, we expect on the sample level that participants with equal probability choose either an Ingroup or Outgroup firm as their bet. However, in the treatment, when they are also making wage decisions at the same time, they may be more likely to bet on the Outgroup firm. This belief that the candidate has worked at a different firm may make the participant feel less similar to the candidate and therefore justify offering a lower wage.
Stage 3 – joint work in the new firm
One of the two candidates will be randomly selected to join the new firm. The task is as in the experiment 1.
Experiment 3: Strategic information avoidance or seeking for strategic belief formation
Stage 1 – gaining job experience
Same as in the experiment 1.
Stage 2 – promotion in a new firm
This stage is very similar to stage two in the previous experiment, with the only difference being that participants can obtain complete the employment history of the recipient at no additional cost.
In the control arm, participants will first state their belief, as described in the previous experiment. Then, on the next page, they have the option to gain or avoid the information about the unobserved firm for which one candidate has worked. This decision does not incur any cost. Subsequently, participants will also state their belief and decide whether they want complete information about the other candidate.
In the treatment arm, participants will first state their belief and choose a wage for a candidate and then have the option to learn the unobserved firm for which a candidate has worked. One of the candidates will be randomly selected, and the chosen wage determines the payoff.
In the treatment, we aim to test if, in the presence of financial incentives, participants will opportunistically gain or avoid information about the employment histories of the candidates to find dissimilarities that would justify their wage decisions.
Stage 3 – joint work in the new firm
One of the two candidates will be randomly selected to join the new firm. The task is as in the experiment 1.