Abstract
When the relative ratio of positive to negative behaviors in a majority and a minority group is constant, illusory correlation originally demonstrated by Hamilton & Gifford (1976) have been shown to create more positive impression of a majority than a minority if the base rate of negative behaviors is low and to create more positive impression of a minority than a majority if the base rate of positive behaviors is low (H G pattern). But this H G pattern has not been always replicated. Most Japanese studies (e. g., Shirai, 1979; Sugimori, 1991) have shown that one group (in most cases, a majority) is consistently perceived as more positive than the other (in most cases, a minority) regardless of the base rate of positive/negative behaviors (S S pattern). To explain these incongruities, we proposed the hypothesis. In the experimental paradigm of illusory correlation, two important pieces of information are given to subjects: behavior of each person and group membership. The difference of subjects' information processing perspectives could have contributed to different patterns of illusory correlation. Our results supported this hypothesis. Under the condition where subjects were more focal to the behavior of each person and the individual-based information processing was promoted, the results replicated the H G pattern. By contrast, under the condition where subjects were more attentive to group membership and the group-based information processing was promoted, the results replicated the S S pattern. These findings suggest that the information processing perspectives take the highly significant roles regarding the illusory correlation phenomena.
