We administered the go-no-go paradigm to 44 boys with attention deficit disorder (ADD) and 32 control subjects who did not have ADD. This task requires a subject to emit a simple motor response to one cue while inhibiting the response in the presence of another cue. Commission errors suggest impulsivity, and omission errors suggest inattention. ADD subjects made more total errors than did control subjects (p less than 0.03), and more ADD subjects made multiple errors (p less than 0.001). Within the ADD group, the nonhyperactive (ADDnoH) subjects were characterized by a high number of commission errors early, and significant improvement with practice (p less than 0.01). In contrast, the hyperactive ADD subjects (ADD/H) did not differ from control subjects in number of early commission errors, but differed from both control subjects and ADDnoH subjects in their failure to improve with practice. In addition, the incidence of omission errors was highest in the ADD/H group. This paradigm can be easily incorporated into the assessment of children with suspected ADD and provides an objective measure of inattention and impulsivity. Our data provide cognitive support for the empirical distinction between hyperactive and nonhyperactive children with ADD.