Background Despite a female advantage in verbal memory, normative data for verbal memory tests used to diagnose Alzheimer’s disease (AD) dementia and amnestic mild cognitive impairment (aMCI) often are not sex-adjusted. Objective To determine whether sex-adjusted norms improve aMCI diagnostic accuracy when accuracy was evaluated by progression to AD dementia over time. Methods Non-sex-specific and sex-specific verbal memory test norms were incorporated into Jak/Bondi aMCI criteria and applied to older (age 65–90) non-demented women (N = 1,036) and men (N = 355) from the Rush Memory and Aging Project. Using sex-specific aMCI diagnosis as the “true” condition versus non-sex-specific aMCI diagnosis as the “predicted” condition, we identified True Positives, False Positives, True Negatives, and False Negatives and compared AD dementia risk over 10 years among groups. Results Rates of aMCI were higher in men versus women (χ 2 = 15.39, p < 0.001) when determined based on typical diagnostic criteria, but this difference reversed when using sex-specific diagnostic criteria (χ 2 = 8.38, p = 0.004). We identified 8%of women as False Negatives and 12%of men as False Positives. Risk of incident AD dementia in False Positive men was significantly lower than in True Positive men (HR = 0.26, 95%CI = 0.12–0.58, p = 0.001). Risk of incident AD dementia in False Negative women was substantially higher than in True Negative women (HR = 3.11, 95%CI = 2.09–4.63, p < 0.001). Conclusion Results suggest that previous reports of higher aMCI rates in men versus women may be an artifact of non-sex-adjusted norms/cut-scores. Incorporation of sex-specific norms/cut-scores for verbal memory impairment into aMCI diagnostic criteria may improve diagnostic accuracy and avoid diagnostic errors in approximately 20%.