Explaining cooperative behavior in public goods games: How preferences and beliefs affect contribution levels

There is a large body of evidence showing that a substantial proportion of people cooperate in public goods games, even if the situation is one-shot and completely anonymous. In the present study, we bring together two major endogenous factors that are known to affect cooperation levels, and in so doing replicate and extend previous empirical research on public goods problems in several important ways. We measure social preferences and concurrently elicit beliefs on the individual level using multiple methods, and at multiple times during the experiment. With this rich set of predictor variables at the individual level, we test how well individual contribution decisions can be accounted for in both a one-shot and a repeated interaction. We show that when heterogeneity in people's preferences and beliefs is taken into consideration, more than 50% of the variance in individual choice behavior can be explained. Furthermore, we show that people do not only update their beliefs in a repeated public goods game, but also that their social preferences change, to some extent, in response to the choices of other decision makers.