The validity of correcting for demographic variables when considering neuropsychological test scores was evaluated in a sample of 136 patients with traumatic brain injury (TBI) who had been screened carefully for premorbid or comorbid confounding factors. When considered in concert with neurological variables, age and education accounted for a significant proportion of the variance in raw scores on the Category Test and the Trail Making Test in the complete sample. Gender did not affect level of test performance. Correcting neuropsychological test scores for demographic variables did not significantly alter their success in identifying patients with severe TBI, but did lead to greater accuracy when classifying individuals with mild-moderate TBI. This investigation concluded that norms that consider the demographic background of the individual are likely to reflect more accurately the neuropsychological status of patients with TBI than interpretations that are based exclusively on raw data.