The domain specificity and generality of mental contamination: accuracy and projection in judgments of mental content.

In this study we examined individual differences across two tasks requiring people to judge the mental contents of the minds of other people--an opinion prediction task and a knowledge prediction task. The tendency to overproject one's own mental contents in both of these tasks has been interpreted as an instance of mental contamination by Wilson and Brekke (1994). Results demonstrate no domain generality in the process of projecting one's own internal states onto predictions about the internal states of others. Furthermore, projection was efficacious in the opinion prediction task but not in the knowledge prediction task. The differing consequences of mental contamination across these tasks was moderated by the presence of other diagnostic cues that were negatively correlated with the diagnosticity of one's own mental contents in the knowledge prediction task but not in the opinion prediction task. Mental contamination was largely unrelated to cognitive ability or to styles of epistemic regulation. However, predictive accuracy (and its primary determinant--use of other diagnostic cues) was correlated across the two tasks, and was also related to cognitive ability and styles of epistemic regulation. The results are interpreted within the context of two-process models of cognitive functioning (e.g. Evans & Over, 1996; Sloman, 1996).