Abstract

Differential performance evaluations among subgroup members were tested in two experiments. Three variables of group membership, evaluation method, and publicity of evaluation were introduced in experiment 1, and group membership and outcome dependency were manipulated to explore any possibility of reducing intergroup bias in experiment 2. Core members' outcome was overestimated and peripheral members' greater deviation was confirmed in both experiments. Also, overestimation of peripheral members was significant in absolute evaluation and private condition in experiment 1. Core members' outcome was significantly biased and greater deviancy noticed among peripheral members in outcome dependence situation. But there was no comparable tendency in outcome independence situation in experiment 2. Overestimation among peripheral members was explained as a getting instrumental basking in the reflected glory by identifying core members. Some conditions, which could reduce intergroup bias when evaluating in/out groups' performance were suggested.