Abstract
Two cohorts of children were followed to determine whether tests of phonological awareness (Syllable Tapping), orthographic processing (Visual Matching), and serial naming speed (RAN Objects), added to a preschool battery, would improve prediction of reading. The major predictors of first-grade reading and spelling were preschool letter naming and sentence memory for both cohorts, but the orthographic and serial naming tasks added a small amount of additional variance. Sentence memory accounted for the most variance in second-grade reading for both cohorts, and Visual Matching made contributions to reading and spelling for each cohort. Sentence memory, Visual Matching, and color naming together yielded an 87% to 90% hit rate in predicting which individual children would be good or poor readers. The orthographic and serial naming speed tasks are useful additions to a preschool predictive battery, but recommendations are that alternative preschool phonological tasks, not based on syllable recognition, should be used to predict reading.