Judgment of neuropsychological decline is typically made by comparing a patient's current cognitive performance to data from demographically similar normal individuals. Even within narrowly defined demographic categories, however, there is variability in level of performance, approximating the normal curve. The present study explored the degree to which oral reading scores on the American National Adult Reading Test (ANART) could more accurately predict a person's test performance relative to other demographically similar individuals. In a sample of 141 neurologically healthy participants, the ANART added modestly to the precision of WAIS–R Verbal and Full Scale IQ and Learning score predictions, beyond that achieved by demographics alone; however, ANART score did not significantly improve estimation of Performance IQ, Average Impairment Rating, or Memory score. Use of the ANART tended to improve demographic predictions primarily with “outlier” participants whose oral reading skills were relatively poor. For Verbal IQ, ANART helped with participants who had both poor ANART and relatively high education. Oral reading can be useful for estimating premorbid verbal intelligence and learning in combination with demographic variables, but it does not appear to improve estimates of other neurocognitive abilities. (JINS, 1999, 5, 247–254.)