The purpose of this investigation was to study the effects of priming positive and negative expectancy outcomes on the drinking responses of college students. Men and women (N = 64) were randomly assigned to 1 of 3 priming conditions: a positive expectancy outcome condition, a negative expectancy outcome condition, and a neutral (control) condition. Participants were exposed to a series of semantic primes corresponding to their condition and then asked to complete a beer taste-rating task. Planned comparisons revealed that the average ratio of beer consumed to body weight in the positive condition was significantly greater than the average ratio in the neutral condition, and the average ratio of beer consumed to body weight was significantly less in the negative condition than the average ratio in the neutral condition. These findings are discussed as they relate to cognitive models of alcohol use.