The relationship of cognitive impairment to the course of schizophrenia remains uncertain. By studying psychotic adolescents, 90% of whom were hospitalized for the first time, we hoped to reduce the influence of such confounding variables as lengthy disease process, neuroleptic treatment, and institutionalization. 39 psychotic adolescent subjects who fulfilled DSM-III criteria for schizophrenia, schizophreniform psychosis, paranoid schizophrenia, or atypical psychosis were compared to 41 non-psychotic adolescent psychiatric controls. Subjects were administered the Wechsler Intelligence Scale for Children-Revised, Peabody Individual Achievement tests of reading, reading comprehension, and mathematics, Bender-Gestalt, and Purdue Pegboard test within 3 weeks of admission to a psychiatric hospital. Performance IQ was significantly lower in the psychotic group (72 versus 93, P = 0.03). Thus, the IQ pattern in adolescent psychotic patients at an early stage in their illness was similar to the pattern displayed by chronic adult schizophrenic patients. Results were not consistent with theories of left hemisphere involvement in schizophrenia. Academic achievement was similar in both groups despite marked differences in performance IQ. Psychotropic medication had no significant impact on the results. In summary, deficits in processing novel material seem at the very least to be present at the onset of the psychotic disorder, though they may be non-progressive thereafter.