Because the diagnosis of dementia is highly dependent on the identification of neuropsychological deficits, there have been increasing concerns regarding cultural bias in tests designed to measure cognitive and intellectual function in English-speaking and Spanish-speaking adults. Despite their widespread clinical use, the effects of potential cultural bias on these measures among these two groups are largely unknown. We assembled a group of 76 females who received an NINCDS-ADRDA clinical diagnosis of Alzheimer's disease (AD). Half of these subjects were primary Spanish speakers, and the other half reported English as their primary language. All subjects were matched on chronological age and severity of memory impairment. Factors such as educational attainment and depression were entered into the analyses as covariates. The two groups were compared with respect to their performance on a battery of neuropsychological tests that are representative of what is typically used to diagnose the presence and severity of dementia. The finding that Spanish-speaking AD patients scored lower on specific neuropsychological measures is discussed in terms of inherent language biases and the possible lack of saliency of a number of these tests. The results obtained further suggest the need to modify certain neuropsychological indices so that they will better serve diverse ethnic and cultural groups.