A neuropsychological screening battery including the Mini-Mental State Examination and four other brief cognitive tests (Russell's Adaptation of the Visual Reproduction Test, Trail Making Test, Verbal Fluency Tests on letters and category, and the Buschke Selective Reminding Test) was administered to a randomly selected population sample of 403 subjects aged 68 to 77 years to evaluate the effect of education, age, and sex on test scores. The difference in neuropsychological screening tests between various education groups (3 years or less, 4 to 6 years, 7 years or more) was statistically highly significant, even after the adjustment for the effect of age. The subscores and total scores were lowest in the minimal education group on every neuropsychological test. Education correlated more strongly than age with all neuropsychological test scores and subscores. The effect of sex on test results was seen only in some subscores of brief neuropsychological tests but not in a single item of the Mini-Mental State Examination. On the basis of our results, the effects of education, age, and sex have to be evaluated before using brief neuropsychological tests in population-based dementia screening.