This study compared three alternative scenarios for effecting a 25% reduction in U.S. alcohol consumption in terms of their respective impacts on the prevalence of alcohol abuse and dependence. The three approaches were (1) an overall 25% reduction in the volume of ethanol intake for all current drinkers, (2) an equivalent absolute reduction taken only among drinkers whose current intake ever exceeds moderate drinking guidelines, and (3) an equivalent reduction taken only among drinkers whose current intake usually exceeds moderate drinking guidelines. The per-occasion cutpoint for moderate consumption was set at the intake level demonstrated to produce psychomotor impairment, and was based on each individual's total body water level. The impact of reducing consumption on the prevalence of alcohol use disorders was estimated by means of a logistic regression model that adjusted for sociodemographic characteristics, family history of alcoholism, and age at first drink, and that took into account interactions between the consumption and other variables. Taking an overall 25% reduction in intake resulted in the same decrease in the prevalence of abuse and dependence (21.7%) as was achieved by taking an equal volume of reduction among only those drinkers whose consumption usually exceeded the moderate drinking cutpoint. Restricting the reduction in consumption to those drinkers whose consumption ever exceeded this cutpoint resulted in a slightly greater reduction in alcohol use disorders, 24.6%.