Identity and group conflict

We experimentally investigate the effects of real and minimal identities on group conflict. In turn we provide a direct empirical test of the hypotheses coined by Amartya Sen that the salience of a real identity escalates conflict but that of a mere classification would not do so. In a baseline treatment, two groups - East Asians and Caucasians - engage in a group contest, but information on the racial composition of the groups is not revealed. In the minimal identity treatment each group is arbitrarily given a different color code, whereas in the real identity treatment the race information is revealed. Supporting Sen׳s hypotheses, we find that compared to the baseline, free-riding declines and conflict effort increases in the real identity treatment but not in the minimal identity treatment. Moreover, this occurs due to an increase in efforts in the real identity treatment by females in both racial groups.