|| Checking for direct PDF access through Ovid
College student drinkers (N = 314) participated in a health survey in which they (a) completed an alcohol-related memory association task (expectancy accessibility measure), (b) rated their positive expectancies about alcohol use (expectancy strength measure), and (c) reported their level of alcohol involvement. Hierarchical regression analyses showed that both expectancy accessibility and expectancy strength predicted frequency of alcohol use and alcohol-related problems. Moreover, moderational analyses showed that the association between expectancy strength and frequency of alcohol use was greater for those who generated more alcohol responses on the expectancy association task. These findings suggest that the outcome association measure and Likert scale ratings of expectancies may assess distinct properties of expectancy representations, which may have independent and interactive effects on different aspects of drinking behavior.