Effects of Dynamic Testing on the Classification of Learning Disabilities: The Predictive and Discriminant Validity of the Swanson-Cognitive Processing Test (S-CPT)

Abstract

This study assessed the criterion-related and discriminant validity of the Swanson-Cognitive Processing Test (S-CPT). The S-CPT is an individually administered standardized battery of information processing tasks that assess performance under initial (static) and dynamic testing (testing-the-limits) conditions. In Experiment 1, children's (N = 61) performance on the S-CPT under initial and dynamic testing conditions was compared to the WISC-R in predicting performance on Reading and Mathematics subtests of the WRAT-R. Gain scores established during dynamic testing conditions were the best predictors of reading and mathematics performance (14% and 26% of the variance, respectively). In Experiment 2, 156 children with reading deficits and 35 skilled readers were compared on the S-CPT under static and dynamic testing conditions. As expected, skilled readers performed better than learning-disabled readers across all subtests, but the results were qualified when subgroups were analyzed. The cluster analysis indicated that approximately 25% of the sample with learning disabilities was misdiagnosed, and better reflected slow learners (poor readers) or instructionally responsive children. The results support the notion that dynamic assessment contributes significant variance in the prediction of academic achievement and that such assessment procedures discriminate the processing potential of children with reading disabilities from children with other types of learning problems.