Knowing what you'll do: effects of analyzing reasons on self-prediction

Analyzing the reasons why one would or would not act in a certain way was predicted to increase the perceived likelihood of the behavior and to lower the accuracy of the self-predictions. In 3 studies, college students predicted whether they would act in friendly or unfriendly ways toward an acquaintance. Those asked to analyze reasons why they would or would not perform the behaviors, as compared with no-analyze controls, were more likely to say they would perform the behaviors, showing a confirmation bias; made less accurate predictions, because analyzing reasons changed their predictions but not their actual behavior; and were more overconfident, because analyzing reasons lowered accuracy but not confidence. Each of these effects was especially pronounced when people's initial liking for the target person was different from the valence of the behavior they were predicting.