INTRODUCTION: The Alcohol Use Disorders Identification Test (AUDIT) is under-studied among young drinkers, as are the contributions of individual items to total AUDIT scores, and online performance of the existing briefer versions of this instrument. DESIGN AND METHODS: This study examined individual items of the AUDIT, and various combinations, including the existing briefer versions, in relation to total AUDIT scores in a Web-based study of young drinkers. A total of 167 young people aged 16-24 years old who had consumed any alcohol within the previous 7 days were recruited by both offline and online methods. RESULTS: Considered individually, items 3, 4, 5 and 8 were predictive of the majority of the variance in total AUDIT scores in this Web-based study. Existing briefer versions of the AUDIT do not better predict total scores than possible alternative combinations of items, for which acceptable levels of sensitivity and specificity for screening have been demonstrated. CONCLUSIONS: Short forms of the AUDIT, particularly those based only on consumption questions, require further validation study in online applications with young people.