Abstract

In two experiments we investigated the cause of the subjects' evaluations of groups in illusory correlation paradigms. In exeperiment 1 there was no difference in the numbers of stimulus informations favored group A or group B. However group A was evaluated more positively than group B. This result showed that the subjects' differential evaluations found in illusory correlation paradigms were not caused by the differences in the numbers of relative informations presented as was assumed in the Woo-Young Chun and Hoon-Koo Lee(1991). The exeperiment 2 showed that subjects' differential evaluations of the two groups were caused by the differences in the numbers of informations representing group which influence the subjects to evaluate the nature of a group as "good" or "bad". Also, as in the results of the previous study of Woo-Young Chun and Hoon-Koo Lee(1991), even in the absence of the ditinctiveness co-occurrence in the encoding stage, the subjects' errors of estimation and recognition errors occured and the differential evaluation of the group was maintained. These results altogether suggest that the evaluations of groups and perceptual errors are not the result of the subjects' overestimations of the correlation between the minority group and infrequent informations in the encoding stage. Implications of the experiments for the formation of group streotypes are discussed.