Aims To test whether an expectancy challenge (EC) changes implicit and explicit alcohol-related cognitions and binge drinking in young heavy drinkers. This is important for theoretical and practical reasons: the EC presents a critical test for the hypothesized mediational role of alcohol cognitions and the EC has been presented as a promising intervention to counter alcohol problems in heavy drinking youth. Setting, participants and intervention Ninety-two heavy drinking college and university students (half women) were assigned randomly to the EC or control condition (a sham alcohol experiment in the same bar-laboratory). Measurements Explicit alcohol cognitions and alcohol use were assessed with paper-and-pencil measures. Alcohol use was assessed prior to the experiment and during a I-month follow-up. Implicit alcohol-related cognitions were assessed with two versions of the Implicit Association Test (IAT). adapted to assess implicit valence and arousal associations with alcohol. Findings and conclusions The EC resulted in decreased explicit positive arousal expectancies in men and women alike. There was some evidence for a differential reduction in implicit arousal associations, but findings depended on the version of the IAT and on the scoring-algorithm used. In men (but not in women) there was a short-lived differential reduction in prospective alcohol use (significant in week 3 of the follow-up). and this reduction was partially mediated by the decrease in explicit positive arousal expectancies. These findings suggest that an EC successfully changes explicit alcohol cognitions and that this may have short-lived beneficial effects in heavy drinking young men.