The purpose of this investigation was two-fold. First, to examine the level of diagnostic accuracy of psychologists when their decisions were subjected to statistical procedures that analyzed group differences and group membership predictions. Second, to examine the sensitivity of the Wisconsin Card Sorting Test (WCST; Heaton, Chelune, Talley, Kay, & Curtiss, 1993) in differentially identifying children who experience ADHD and RD-LD symptoms. The diagnostic battery used to identify ADHD, RD-LD, and Normal subjects included intellectual, academic achievement, attention, and hyperactivity measures recognized as sensitive to these disorders. Performance on the WCST was then examined to determine its usefulness in discriminating between the aforementioned groups.Scores for the diagnostic variables for 115 subjects (mean age = 9.8 years; males = 80; females = 35) were analyzed using oneway ANOVAs to determine differences between groups. A subsequent cluster analysis was conducted using Ward's method to determine group membership of the subjects and resulted in a sample of 87. This cluster analysis resulted in a four cluster solution with the groups being identified as ADHD, RD-LD, Normal, and "Close Calls."Two linear discriminant analyses were performed with the first using the diagnostic groups diagnosed by the previous psychologists as groups and diagnostic variables as predictors. The second used the diagnostic groups diagnosed by evaluating psychologists and the WCST variables used as predictors to examine their ability to discriminate between groups and predict membership.The first linear discriminant analysis yielded two significant functions of three indicating confidence in the diagnoses provided by the evaluating psychologists. The second linear discriminant analysis yielded no significant findings when using the WCST variables as predictors. As a result, it appears the WCST provides little useful information in the differentiation between ADHD, RD-LD, and normals. / Department of Educational Psychology