Maintaining Consistency between Self-Serving Beliefs and Available Data

It was hypothesized that individuals evaluate data relevant to outcome attributions in a manner that enables them to maintain logical consistency between the available evidence and their self-serving attributions for the outcome. Subjects were led to succeed or fail on a bogus social sensitivity test and then were given information concerning two studies, one of which concluded that the test was valid and the other that the test was not valid. As predicted, success subjects evaluated the high-validity conclusion study more favorably and the low-validity conclusion study less favorably than did failure subjects. Furthermore, exposure to the mixed evidence after the performance feedback led to increased selt-ratings of social sensitivity among success subjects, but had no effect on failure subjects. The implications of these results for understanding how individuals generate and maintain self-serving beliefs were discussed.