Excessive alcohol consumption among college students is a serious problem. Alcohol education courses, during which students are taught to free‐pour accurate servings of alcohol, have been proposed as one strategy to reduce this problem. However, few studies have evaluated the effects of peer models on free‐pour skill maintenance. We used a nonconcurrent multiple‐baseline across subjects design to assess the effects of peer modeling on maintenance of college students' (N = 19) pouring skills of a standard serving of beer immediately following Behavioral Skills Training (BST). Results indicated participants who inaccurately poured a standard serving of beer at baseline (n = 17) accurately poured following receipt of BST, and all participants (N = 19) maintained accurate pouring in the presence of peer confederate models who poured inaccurately or accurately. Results suggest BST is an efficient (i.e., low to no‐cost, requires no sophisticated laboratory space or software, and makes use of available and affordable materials) and effective tool for teaching college students to pour standard servings of beer and that the immediate introduction of inaccurately pouring peer models has little or no effect on skill maintenance.