Two experiments yielded further evidence for the black sheep effect (Marques, Yzerbyt and by ens, 1988). In the first experiment, 66 subjects were presented with two good or two poor speeches, one supposedly made by an ingroup member and the other supposedly made by an outgroup member. In the second experiment, 37 subjects were presented with one good and one poor speech supposedly made either by two ingroup members or by two outgroup members. The black sheep effect was predicted and found in both experiments: subjects over-evaluated likeable ingroup members and under-evaluated unlikeable ingroup members as compared to equally likeable and unlikeable outgroup members. Collapsing the data of the two experiments suggests that social comparison may be performed, in purely symbolic terms, against a cognitive standard of positivity rather than an outgroup present in the judgmental situation. The emergence of the predicted effect when strongly individualized information was presented in inter- as well as in intra-group situations supports the robustness of the black sheep effect.