Preprints of the Max Planck Institute for Research on Collective Goods 2014/19
In a laboratory experiment, I test whether guilt aversion, i.e., a preference to fulfill other people's expectations, plays out stronger if agents are socially close. I induce two different minimal group identities in participants and randomly assign participants to one of two treatments. Senders either play a dictator game with a receiver from their own group (ingroup treatment) or from the other group (outgroup treatment). I let senders condition their amount sent on second-order beliefs. I find that, in the realm of realistic beliefs (i.e., the sender expects the receiver to expect the sender to send no more than half of the pie), the positive influence of second-order beliefs on how much the sender sends is stronger in the ingroup treatment. In both treatments, about half of the senders remain unaffected by second-order beliefs. In the ingroup treatment, unaffected senders identify less with their group than affected senders do. This is not true for the outgroup treatment.