|| Checking for direct PDF access through Ovid
A growing body of research has examined self-reported measures of social network alcohol use, such as the Important People Instrument (IPI), among college students. However, it remains unknown whether IPI judgments of friends’ alcohol use are accurate. We hypothesized that judgments of friends’ drinking status (e.g., heavy drinker) and maximum drinks per day would primarily reflect friends’ self-report (accuracy), rather than projection of participants’ self-report (assumed similarity) or a systematic tendency to under- or overestimate behavior (directional bias). We also expected that accuracy would be stronger when participants were women, had more contact with friends, or drank less each week or when friends drank less each week. In all, 654 randomly selected, 1st-year students indicated their 5 closest friends in their class, yielding 111 friendship dyads. Participants judged each friend’s drinking status and maximum drinks per day and rated these items for themselves. Gender, frequency of contact, and typical drinks per week were assessed. Results indicated that judgments of drinking status and maximum drinks per day were highly accurate. Accuracy effects were consistently stronger than was assumed similarity; directional bias was nonsignificant. Accuracy did not depend on gender or participants’ weekly consumption but was stronger for both outcomes when contact was more frequent and, for maximum drinks per day, when friends’ weekly consumption was relatively low. Results support validity of the IPI for assessing social network alcohol use among students. Given that perceptions are accurate, research is needed on intervention strategies that facilitate management of risky peers.