Thirty children with subaverage IQs and psychiatric diagnoses of attention deficit disorder and/or or conduct disorder took part in a double-blind study of placebo, methylphenidate, and thioridazine, which were given for 3 weeks each. The results showed a consistent and highly significant effect of methylphenidate in reducing teacher ratings of problem behavior. Parent ratings showed no behavioral effects for the group as a whole. An attentional model of stimulant drug response was used to divide subjects according to a cognitive maturity domain presumed to reflect selective attention. When divided according to breadth of attention, mental age, and IQ level, higher functioning subjects were found to show a generally favorable response to methylphenidate on both teacher and parent rating scales, whereas children of low functional level typically showed an adverse or indifferent response. The present data suggest that mental age and IQ may be important determinants of drug response; below a given level, there was a greatly reduced likelihood of responding positively. Clinical response to thioridazine was substantially less than the response to methylphenidate, with significant improvements confined to conduct and hyperactivity problems on teacher ratings.