Using data from the Early Childhood Longitudinal Study Kindergarten Cohort of 1998 (n = 15,070), this study used propensity scores to examine the short- and long-term academic and psychosocial benefits of preschool education for a diverse sample of middle-class children. Compared with children who attended informal care at age 4, preschool attendees consistently performed better on achievement tests from age 5 through early adolescence, but exhibited less optimal psychosocial skills. These negative behavioral effects of preschool were concentrated among children who attended preschool for 20 or more hours per week, but otherwise, there was little evidence of heterogeneity as a function of program type or child- and family-characteristics. The long-term academic advantages of preschool were, however, largely explained by their positive effects on academic skills early in formal schooling and there was evidence for convergence in children's academic test scores, which was partially attributed to the differences in children's social skills during the early elementary school years. [This is the online version of an article published in "Journal of Educational Psychology."]