Gordon Diagnostic System and WISC-III scores for clinic-referred 6- to 16-yr.-olds (184 with ADHD Combined Type and 46 without ADHD) were analyzed to evaluate the combination of scores and cutpoints that maximized diagnostic accuracy. Using an "ADHD cutpoint" of IQ minus the GDS Composite score > or = 13, 87.8% of the children were correctly identified as having or not having ADHD. IQ minus Freedom from Distractibility > 0 yielded 73.5% accuracy. When the two meaures were combined to create new criteria (IQ minus GDS Composite > or = 13 or IQ minus Freedom from Distractibility > or = 11), diagnostic accuracy increased to 90.9% and negative predictive power improved substantially. Diagnostic agreement between the Gordon Diagnostic System and Freedom from Distractibility was 70%, suggesting that the two tests measure both similar and unique traits.