The Perry Preschool program was a randomized social experiment with long run followup that supplemented the early environments of disadvantaged African American children. Evidence from the program is widely used in support of claims of the effectiveness of early childhood intervention programs. The Perry program has been shown to boost adult performance for both girls and boys. It has a rate of return above the historical return to equity. The program had little lasting impact on the IQs of its participants. Using newly discovered data on the noncognitive traits of treatments and controls collected measured a few years after the program finished, we find that experimentally-induced changes in noncognitive traits explain a sizable portion of later-life treatment effects for education, employment, earnings, and crime. This paper contributes to an emerging literature on the importance of noncognitive traits in explaining a variety of social outcomes.