This study tested hypotheses derived from a theoretical model in which interests are a product of a person's modified cognitions of his or her perceived ability to perform, expected success in, and anticipated satisfaction from engaging in an activity. In a field experiment with 3 groups of preschool children (n = 90), preferences for unfamiliar game activities were either cognitively restructured in line with the model's cognitive factors, rewarded through external reinforcers, or not manipulated. Participants were asked to rank their attraction to the activities before the manipulation, immediately after their first (restructured, rewarded, or unmanipulated) exposure to four activities, and 2 weeks later. Results for both genders provided clear support for the argument of cognitive modification of interests. The differential development of interests is discussed from a cognitive-behavioral perspective and in relation to Bandura's self-efficacy theory.