Does Shared Group Membership Promote Altruism?

Abstract
Two explanations of why shared group membership promotes cooperation in social dilemmas were compared. According to the fear—greed model of social identity proposed by Simpson (2006), shared group membership reduces greed but not fear and, thus, should promote altruistic behavior toward in-group members in the absence of fear. According to the group heuristic model proposed by Yamagishi and colleagues, altruistic behavior toward in-group members is a `ticket' to enter a generalized exchange system; people are not predicted to behave altruistically when it is made salient that no system of generalized exchange operates in the group. We tested these models in a dictator game experiment with two conditions. In the common knowledge condition, either model predicts greater altruism toward in-group recipients. In the unilateral knowledge condition — when the dictator knows the group membership of the recipient, but the recipient does not know the membership of the dictator — the fear—greed model predicts greater altruism toward ingroup recipients. In contrast, the group heuristic model does not predict such in-group bias in altruism. The results of the experiment support the group heuristic model.