The study reported here sought to establish the predictive validity of the Miller Assessment for Preschoolers (MAP), an instrument designed to identify preschool children at risk for school-related problems in the primary years. Children (N=338) in 11 states who were originally tested in 1980 as part of the MAP standardization project were given a 4-hour battery of standardized tests. Standardized criterion measures included the Wechsler Intelligence Scale-Revised, the Berry Developmental Test of Visual Motor Integration, the Bruininks-Oseretsky Test of Motor Proficiency, the Goodenough-Harris Drawing Test, the Woodcock-Johnson Psychoeducational Battery, and the Walker Problem Behavior Identification Checklist. Parents and teachers completed behavior checklists and history forms. Among subjective criterion measures were retention in school; teacher observations; receipt of special services; and report card grades in language, reading, mathematics, and physical education. Predictor variables examined were the MAP Total Score, the five MAP subtest scores, and the 27 MAP items. Data were subjected to correlational, T-test, stepwise multiple regression, and classificational analyses. Detailed results indicate that the predictive validity of the MAP compares favorably to similar predictive validity studies, and that the instrument can consistently and correctly identify the majority of children who later have difficulty in school, with relatively few misclassifications. (JW)