Based on theoretical considerations drawn from John Dewey and others, and using the Experience Sampling Method to longitudinally investigate a group of talented high school students, undivided interest was operationalized as times when the students felt above average spontaneous interest (i.e., excitement, openness, and involvement) while also reporting above average goal-directed interest (i.e., that their task was important to their goals). Results showed that after adjusting for the effects of family background, scholastic aptitude, and other individual differences, undivided interest while doing talent-related activities was positively correlated with independent assessments of talent area performance three years later: the level of mastery students achieved as indicated by their school records, the ratings students received from their talent area teachers, and the students' assessments of their own level of engagement. Highly engaged students reported over twice as much undivided interest in comparison to a group of disengaged students, who reported more divided interest (i.e., more of what Dewey referred to as ''fooling''-high spontaneous involvement with no goal direction; and more ''drudgery''-low spontaneous involvement and high goal direction). These findings held regardless of whether the teenagers were talented in math, science, music, or art. The implications of the study are discussed in terms of contemporary theories of attention and cognitive development, as well as unproductive educational philosophies that pit these important dimensions of experience against each other.