The present study investigated the differential predictive validity of the Initial Skills Analysis. In the fall, these reading screening tests were administered to students in Grades K to 3. In the spring, a series of reading outcomes tests were administered to these same students, and teachers provided judgments of the students' reading skills. For the male-female, Hispanic/non-Hispanic, and Black/non-Black comparisons, the hierarchical regression models showed that if a single prediction model was used, the screening tests over predicted minority performance at several grade levels. However, no consistent evidence of differential prediction with practical importance was found for any of the groups studied. Although the results were consistent with past research, over prediction of the criterion scores by a screening test could result in minority students not receiving needed remediation services. As a result, further evidence for the validity of these screening tests should continue to be collected.