The National Adult Reading Test (NART) has achieved popularity as a measure of pre-morbid intellectual ability, based on the premises that pronunciation of irregular words is unaffected in many clinical disorders and that performance is highly correlated with general intellectual ability. Recently, schizophrenia research studies have begun to appear in the literature, where the NART has been used to estimate pre-morbid ability. However, this use has preceded the basic required demonstration that, in fact, NART performance is unaffected by the schizophrenic process. In the present study, NART performance was compared across three groups; 20 acutely ill unmedicated DSM-IIIR schizophrenics, 10 other unmedicated acute psychotics, and 20 control subjects. When demographic variability between the groups was controlled for, there were no group differences in terms of NART performance. NART performance was not correlated with Brief Psychiatric Rating Scale scores, and in all three groups, no significant differences emerged when demographically predicted intelligence quotients were compared with NART estimated intelligence quotients. NART performance (predicted on the basis of demographic variables) was not significantly different from observed NART performance in any of the three experimental groups. However, within the sample with schizophrenia, NART estimated pre-morbid IQ was significantly higher than currently measured intellectual abilities. These results suggest that the National Adult Reading Test provides a reasonable estimate of pre-morbid ability in acutely ill, unmedicated schizophrenic patients.