Since its publication in 1982, the National Adult Reading Test (NART; Revised Version, NART–R) has become a widely accepted method for estimating premorbid levels of intelligence in neuropsychological research. However, the assumption that NART/NART–R performance is relatively independent of brain damage has been increasingly challenged in recent years. In a number of conditions, including Alzheimer dementia and Korsakoff's syndrome, studies have indicated a deterioration in reading ability, leading to an underestimated premorbid IQ. In a reaction to these studies, some researchers have advocated the use of demographic variables as a more suitable foundation for accurately predicting premorbid intelligence. We addressed this issue by calculating IQ estimates on the basis of NART/NART–R, demographic variables, and a combination of the two approaches and by comparing these with current WAIS/WAIS–R IQ in patients with Korsakoff's syndrome, Alzheimer dementia, frontal or temporal lobe lesions, and in healthy controls. Estimated premorbid IQs did not differ across groups, whether derived from NART/NART–R or demographic variables. Those based on NART/NART–R demonstrated higher correlations with current WAIS/WAIS–R IQ in controls and patients than those derived from demographic variables. An equation combining NART scores with demographic variables did not significantly increase the amount of variance in IQ explained by NART only, either in patients or controls. The data offer reassurance regarding the continued use of NART as a valid estimate of premorbid intelligence in a number of conditions. (JINS, 2002, 8, 847–854.)