Predicting behavior in economic games by looking through the eyes of the players

Social scientists often rely on economic experiments such as ultimatum and dictator games to understand human cooperation. Systematic deviations from economic predictions have inspired broader conceptions of self-interest that incorporate concerns for fairness. Yet no framework can describe all of the major results. We take a different approach by asking players directly about their self-interest--defined as what they want to do (pleasure-maximizing options). We also ask players directly about their sense of fairness--defined as what they think they ought to do (fairness-maximizing options). Player-defined measures of self-interest and fairness predict (a) the majority of ultimatum-game and dictator-game offers, (b) ultimatum-game rejections, (c) exiting behavior (i.e., escaping social expectations to cooperate) in the dictator game, and (d) who cooperates more after a positive mood induction. Adopting the players' perspectives of self-interest and fairness permits better predictions about who cooperates, why they cooperate, and when they punish noncooperators.