Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Two prevailing theories exist to explain why people engage in costly punishment in one-shot interactions when they cannot expect to gain from the punishment or to affect future behavior: to achieve balance or to send a message. In this paper we constructed new versions of the Dictator, Public Good, and Power-to-Take games, designed specifically to allow us to differentiate between these explanations. The results lend some support to each of the hypotheses; on the one hand, subjects punish even when the punisher knows that the offender will never learn that he was punished, and, on the other hand, there is more punishment when the punisher knows that the offender will be informed that he is being punished than when the punisher knows that the offender will remain oblivious to the punishment. Thus, punishers seek both "retributive justice" and to "send a message" to the offending party.