Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Teams are known to behave differently from individuals, but whether they behave more cooperatively or selfishly is still unsettled in the literature. We let subjects form two-person pairs and play a finitely-repeated two-player public goods game with other pairs, and then compare the pairs’ behaviour with the behaviour of individuals in the same game played against individuals. We vary treatments by the matching protocol between decision-making units (partner or stranger matching). Our data show that when the matching is fixed for all periods, pairs are able to sustain cooperation at high levels while individuals steadily decrease contributions from period to period. By sharp contrast, when pairs are randomly matched with other pairs in every period, they quickly decrease contributions over the periods, as is the case for individuals with the stranger matching protocol.