1997 年 12 巻 3 号 p. 190-198
The purpose of this study is to test a new hypothesis for the group identity effects in social dilemmas, the effect that common group membership promotes cooperation. According to the proposed hypothesis, people prefer to cooperate with ingroup members because they expect reciprocal responses from ingroup, but not from outgroup members. In other words, people are considered to expect generalized reciprocity to exist within groups but not beyond group boundaries. Based on this hypothesis, it was predicted that the previously observed group identity effect-subjects facing a prisoner's dilemma cooperate more with an ingroup member than with an outgroup member-exists only when the partner shares the membership information. When the partner does not know that the subject shares the same group membership, group identity effect is predicted not to emerge. Result of an experiment with 78 subjects clearly supported this prediction and rejected alternative hypotheses based on psychological distance(Krammer & Brewer, 1984), social identity (Billig & Tajfel, 1973), and ingroup stereotype (Brewer, 1978).