This study attempts to determine whether cognitive or sociodemographic/alcohol-related variables were predictive of learning in educational treatment. Eighty-seven male alcoholics were exposed to one hour of instruction on the medical effects of alcohol. Knowledge was assessed 24 hours prior to and 24 hours and three weeks after the intervention. Statistically significant increases in knowledge were found both 24 hours and three weeks following the educational intervention; retention of information declined significantly from 24 hours to three weeks post-intervention. The findings suggest that the clinician may be able to employ a limited number of variables to differentiate between patients who will and will not be able to acquire knowledge from educational interventions. 2 tables, 20 references. (Author abstract modified)