Abstract To delineate putative cognitive effects of estrogen in women with Alzheimer's disease, we compared neuropsychological performances in three groups of patients with clinically diagnosed Alzheimer's disease: women receiving estrogen replacement therapy ( n = 9), women not receiving replacement therapy ( n = 27), and men ( n = 26). Untreated women and men were matched by age, education, and duration of dementia symptoms to women receiving estrogen replacement. We hypothesized that treated women would have better scores on neuropsychological tasks. Results showed that women receiving hormonal therapy performed significantly better than other women on some, but not all, tasks; on no task did women receiving estrogen score significantly worse. The largest group difference was on the Boston Naming Test, a semantic memory task previously shown to be more impaired in women with Alzheimer's disease than in men with this diagnosis. Of tests considered in a discriminant analysis, the naming task was the only neuropsychological variable to distinguish between the two women's groups. Mean differences between estrogen-treated women and men were small and were not statistically significant. Findings support the hypothesis that estrogen therapy for women with Alzheimer's disease is associated with better cognitive skills and that previously noted gender-associated differences in Alzheimer's disease may reflect a state of acquired estrogen deficiency among women with this disorder.