| Abstract: | Distinctiveness-based illusory correlation effects (ICEs) have been implicated in the formation of negative stereotypes of minorities. In standard experiments subjects are presented with information about a large and small group in which the ratio of desirable to undesirable members is the same. These conditions normally lead to a negative representation of the smaller group. Explanations of this effect suggest it is a product of (a) over-representing highly distinctive stimuli, (b) sampling from memory, or (c) regression to the mean following information loss. The present research investigated the possibility that the ICE would occur even where no stimulus information linking behaviours to groups was presented—an outcome inconsistent with the first two explanations. Two experiments (Ns = 83, 112) produced ICEs in the absence of this standard stimulus information in a range of conditions. Results in these and previous studies are shown to be consistent with an explanation in terms of the principle of meta-contrast derived from self-categorization theory. |
