Do choices feed back into and alter preferences? Widespread evidence arising in psychology and neuroscience shows that preferences change in response to own choices, a phenomenon typically explained through cognitive dissonance. The evidence, however, presents serious shortcomings casting doubts on its relevance for economics. We present two experiments addressing these shortcomings. First, participants made standard decisions under risk rather than facing unfamiliar alternatives. Second, all choices were incentivized. Third, our novel experimental design avoids recently-exposed problems of experiments in psychology. The results show unsystematic effects which differ from and challenge conventional wisdom outside economics.