A study by Wiens, Bryan, and Crossen (1993) suggests the Wide Range Achievement Test-Revised (WRAT-R) Reading subtest and North American Adult Reading Test (NAART) are adequate predictors of Wechsler Adult Intelligence Scale-Revised (WAIS-R) IQ scores for a normal population. Although it is common practice to use reading scores to estimate premorbid IQ in clinical populations, the WRAT-R and NAART have not been compared using individuals with brain dysfunction. The current study cross-validated the Wiens et al. (1993) study using neurologically impaired populations: traumatic brain injury (n = 118), dementia (n = 37), and other neurologic impairments (n = 77). The results were generally consistent across all three groups: (a) the WRAT-R and NAART were equivalent and accurate estimates of average VIQ levels; (b) the WRAT-R and NAART were equivalent but underestimates of higher intelligence ranges; and (c) the WRAT-R is a more accurate estimate for lower VIQ ranges, although both are overestimates. This third finding is in contrast to Wiens et al.'s (1993) results that suggest the WRAT-R is an accurate estimate of lower IQ ranges for normals. It is concluded that the WRAT-R is the preferred measure of premorbid verbal intelligence for psychometric and clinical reasons.