LAUSR

In a series of laboratory experiments, two types of players were created randomly. Participants of one type were selected for a group based on performance on a task, whereas participants of the other type were selected automatically without prerequisite. In the main experiment, such favoritism induced a decline in cooperation, measured as contributions in pairwise public goods games, compared to when all participants were treated equally. The reduction in cooperation was observed both for those participants who did not benefit from the favoritism and for those who did, and regardless of whether a player was matched with someone who was favored or not. In extensions of the original experiment, the main results were replicated. Furthermore, the negative effect on cooperation was shown to exist also continue when a rationale was given for the use of favoritism, but to be turned off when selection was random instead of performance-based.