The revised Malmö modification of the Michigan Alcoholism Screening Test (Mm-MAST) of 10 items including all four CAGE (an acronym for Cut down, Annoyed, Guilty, Eye-opener) items and one item on total sobriety was used in a health screening of a total birth-year cohort in Malmö. Seventy-two percent of the female and 68% of the male population participated. Nine hundred and eleven women and 911 men were included in the study population. The revised Mm-MAST yielded a higher proportion of scores indicative of problem drinking than the CAGE in both sexes; 4% of the women and 17% of the men had positive Mm-MAST scores (> or = 3 yes-answers). Rates of positive CAGE scores (> or = 2 yes-answers) were 1% and 7%, respectively. Frequencies of problem drinking in the Mm-MAST and CAGE positive scorers, respectively were similar for both women and men. The reliability of the revised Mm-MAST was better in the male population producing a Cronbach's alpha coefficient of 0.69 compared with 0.58 in the female population. In men the four CAGE items had the highest item-total scale correlations, which was found in only one CAGE item in women. We suggest that the revised Mm-MAST is a better screening instrument than the CAGE in both sexes. Nevertheless the reliability of the total revised Mm-MAST was poorer in women and the need for a female oriented alcohol questionnaire is promoted.