To determine whether social contact (forced versus voluntary) and reward acquisition (winning versus not winning) were differentially effective in influencing positive attitude change toward educable mentally retarded children (EMRs), 68 nonEMR males in the fourth through sixth grades were asked to select either a same-sex EMR or nonEMR as a partner for a bean-bag toss game to help them win a prize. Subjects were able to select the EMR voluntarily or were forced to do so by the experimenter. The game was rigged so the experimenter was able to manipulate winning and not winning the game. Baseline attitude data was collected two weeks prior to the experimental task (T-1), immediately following the task (T-2) and two weeks later (T-3). The results indicated that reward acquisition was more effective than social contact on improving T-2 attitudes scores, but that voluntary social contact was more effective in raising T-3 scores. The findings were discussed in terms of the desirability of integrating EMRs with nonEMRs. (Author)