Used 172 undergraduates in 3 experiments to assess the effects of making salient either a superordinate (collective) or subordinate (differentiating) group identity in heterogeneous groups. In Exp I, 22 male and 36 female Ss were assigned to either a superordinate-group identity (small community resident behavior vs other areas) or a subordinate-group identity (behavior of young people vs elderly people) condition and were asked to perform a computer task individually; Ss were led to believe they were interacting with 5 other persons (2 real and 3 bogus Ss) in their group in accumulating as many points as possible while making the resource last as long as possible. Bogus feedback about group behavior was given. In Exp II, 29 male and 19 female Ss were told that the bogus Ss were economics majors and were asked to perform as in Exp I. In Exp III, the level of social-group identity for 40 male and 26 female Ss was manipulated by varying the common fate of the group members. Results of all 3 experiments show support for the hypothesis that individual restraint would be most likely when a superordinate group identity was made salient and under conditions in which feedback indicated that the common resource was being depleted. A sex-response difference found in Exp I was not sustained in subsequent experiments.
*Reprinted in J. Levine & R. Moreland (Eds.), Key Readings in Group Processes, Psychology Press.