Generalizing the Illusory Correlation Effect

We used two experiments to examine the influence of one's own attitude on the perception of group attitudes. In the first experiment, subjects viewed opinion statements, supposedly made by residents of two towns, on the issue of building a local nuclear power station. One town was large and had frequently occurring statements and the other was small with infrequently occurring statements; there was an equal proportion of pro and anti statements in both towns. The prediction that subjects would perceive an illusory correlation between attitude positions similar to their own (self-relevance) and the infrequently cited (distinctive) town was supported for anti subjects only. Subsequent investigation indicated that this was due to the confounding effect of a prior expectation associating small towns with more antinuclear attitudes. Experiment 2 eliminated the variable of town size by informing subjects that towns of equal size had been more heavily or lightly sampled. Consistent with the hypotheses, both pro and anti subjects perceived an illusory correlation between their own attitude and the town providing the smaller sample, this effect increasing with attitude extremity. The consequences of these findings for the generalizability of illusory correlation explanations of stereotyping are discussed.