The diagnostic utility of the Woodcock-Johnson III Tests of Cognitive Abilities and Clinical Clusters was assessed in a sample of 52 children (26 Attention Deficit/Hyperactivity Disordered (ADHD) and 26 matched controls). Multivariate analysis of variance followed by post-hoc testing and d-ratios yielded some statistically significant and clinically meaningful differences between groups on the Cognitive Fluency Cluster and the Tests of Auditory Attention, and Rapid Picture Naming. Discriminant function analyses indicated that the WJ III COG Tests collectively classified 80.77% of the sample correctly (76.92% of controls and 84.62% of children with ADHD correctly identified). The Auditory Attention and Rapid Picture Naming tests were found to make the most significant contribution overall to the discriminant function. Using a cut-score of 85, the WJ-III COG Clinical clusters and subtests examined in this study offered fair to weak diagnostic utility based on indices of sensitivity, specificity, positive and negative predictive power, as well as results of Receiver Operating Characteristic curve analyses. Implications for research and practice are outlined. [Note: The page range (37-49) shown on the PDF is incorrect. The correct page range for this article is p36-49.]