In Reply.—We thank Glascoe and Hoffman for their thoughtful comments. Our study examined the relationship between tests of developmental and preacademic skills at age 5 years and school achievement at age 8 years. When children did poorly at age 5 years there was only a slightly better than 50/50 chance they would do poorly at age 8 years. Thus, we agree that, based on these findings, it is inappropriate to use prekindergarten tests to predict academic achievement over long intervals.
Examining the predictive validity of preschool tests was not a "methodological weakness" on our part but the purpose of the study. There is great interest in early recognition and remediation of developmental difficulties. In infancy, this appears feasible for moderate to severe developmental delays and deviations. For older children, however, it is more difficult to recognize subtle problems with learning skills until the age at which they typically emerge. Nevertheless, educators