The predictive validity of an original Piagetian instrument (the Piagetian Task Instrument) was compared with that of two frequently used published screening tests for children entering kindergarten: (1) the Brigance K-1 Screen, and (2) the Merrill Language Screening Test. The relative contribution of each test was determined when they were combined with a battery for predicting young children's initial academic success in kindergarten. Subjects were 95 predominantly white, middle-class children (47 boys and 48 girls) without history of developmental difficulties. Subjects' mean age was 60 months. The criterion measure, administered 8 months later, was a teacher's rating scale. Correlation and multiple regression techniques were used to analyze the data. Contrary to the hypothesis, the Piagetian instrument had the lowest predictive validity of the three and contributed least to the regression equation. The predictive validity of the battery was judged too low for use in making decisions about very young children. It is suggested that it is inadvisable to test young children routinely prior to entry into kindergarten. The construct validity of the criterion measure and the predictor measures was discussed. Although the Piagetian test did not display significant predictive capacity, the theoretical base appears to warrant further investigation. (Author/SLD)