An experiment on how wage discretion affects distribution preferences

We conduct our experiment on the online platform Prolific, recruiting participants to a study called ``Decide about final wages to participants in another Prolific study". Participants in this study decide in fact about wages for a previous study, conducted on the same platform which is registered as part of this trial. We randomize participants into the treatment and control group according to the order in which they enter the experiment, where even incoming participants are assigned to be in the control group and odd incoming participants are assigned to be in the treatment group. Participants are asked to decide how to split a bonus of 2 pounds between two pairs of workers who did the same task. We implement these choices for one of the two worker pairs with 50%.

2021-10-20

2021-10-25

The choice of final earnings.

The outcome is participants' choice on how to split the bonus between two workers. They are asked this in terms of percentages. This means that our outcome is the percentage of the bonus awarded to the participant who chose their own share of the bonus and their counterpart in the control group.


Change and explanation. Change decided on 19/10/21 before launching data collection for spectators but after collecting worker data that the spectators will decide on. We had pre-registered rounding to the next 5% in order to be able to create treatment and control group worker pairs that are perfectly matched on performance. After we collected worker data, we found that this a) is not necessary to obtain perfect matches on performance and b) will give an overhang of worker pairs with 50/50 performances, meaning that spectators loose information.
Therefore, we will only round performances to the next 1%, rather than the next 5% before matching.

We conduct our experiment on the online platform Prolific, recruiting participants to a study called ``Decide about final wages to participants in another Prolific study". Participants in this study decide in fact about wages for a previous study, conducted on the same platform. We randomize participants into the treatment and control group according to the order in which they enter the experiment, where even incoming participants are assigned to be in the control group and odd incoming participants are assigned to be in the treatment group. Participants are asked to decide how to split a bonus of 2 pounds between two pairs of workers who did the same task. We implement these choices for one of the two worker pairs with 50%.

Treatment spectators are shown worker pairs where a worker within the pair has decided their own share of the bonus, while control spectators are shown similar worker pairs where a third party made this decision. Thus, spectators in the control group and spectators in the treatment group look at worker pairs with similar performance levels, where either a randomly drawn worker from the pair (treatment group) or a third party worker drawn from the matched treatment group pair (control group) suggests preliminary wages.

The experiment has the following stages (here with an example worker pair performing 60% and 40%):
- Introduction: Welcome to this economics experiment developed at the University of Oslo for research purposes. In this study we ask you about your general opinions about fair pay-offs. You will also decide about the pays-off of participants in an earlier experiment and we will pay out in line with your decisions.
In economics experiments deception is never used. This means that any information you are provided with in the experiment is correct. All interactions are anonymous and are run on a secured server. The administrators of the experiment do not observe your decisions during the whole experiment. At the end of the study, you will be asked to click on a link redirecting you back to the Prolific page. It is very important that you click on this link to prove that you have terminated the study. Remember that if you do close your browser or leave the experiment, you will not be able to re-enter the experiment and we will not be able to pay you!

- Info about experiment: This part of the experiment is not a hypothetical survey. The choices you make in this experiment have real life consequences, because we will pay participants as you decide. We ask you to decide about pay-offs for 2 pairs of participants in a study we ran on Prolific on x. Participants individually solved real-effort tasks for 5 minutes and were then randomly placed into pairs. Every participant receives a fixed participation fee. Additionally, we have reserved a bonus of 2£ for each pair of participants. You will decide how to split this money between them. We select one of your two decisions for payment. For the selected decision, we will flip a coin. With a 50% chance, we will pay the participants exactly as you decide.

2. Inform about preliminary earnings:
We inform you about the relative productivity in the pair of participants. Both participants get a value in percent and the values together add up to 100%. One participant performed at 60%, we call this participant “A”. The other participant performed at 40%, we call this participant “B”. Status quo for payments - this will be paid out if you do not make a decision or if your decision is not selected.
Treatment: We let one of the Prolific participants working on the questions choose preliminary pay-offs. That is, they get to choose both how much to give to themselves and how much to give to the other participant. In this pair, participant A got to decide and made the following choice: "Give 80\% of the total pay-off to myself (participant A, right side) and give 20% of the pay-off to participant B (left side)."
Control: We let a another Prolific participant choose preliminary pay-offs. That is, they get to choose both how much to give to participant A and how much to give to participant B. This third-party made the following choice: "Give 80% of the total pay-off to participant A (right side) and give 20% of the pay-off to participant B (left side)."
3. Control questions
4. Decision 1.
5-7. Repeat stages 2-4 with another worker pair.
8. Open text question asking spectators to reason their choice
9. Background questions.

By order in which participants enter the experiment. Even incoming participants are assigned to be in the control group and odd incoming participants are assigned to be in the treatment group.

Participant in Prolific

No

No clusters

2058 These are the number of participants that we will recruit for our spectator surveys. Part of the intervention is the worker experiment, where the initial earnings choices will be made that the spectators later decide on. In the worker experiment, there will be 2058 deciders in 2058 worker pairs (4114 in total), the same amount as spectators, since we will have one spectator per worker pair.

1029 in the treatment group and 1029 in the control group.