Alcohol use patterns vary significantly within individuals based on drinking context – including location, motives, and social conditions – which directly influence drinking behavior, consumption levels, and alcohol-related consequences. While cross-sectional surveys have examined drinking contexts, intensive longitudinal sampling methodologies such as ecological momentary assessment (EMA) are essential for capturing proximal predictors of heavy episodic drinking. Although studies have established associations between contextual factors and alcohol outcomes, the brief assessment requirements in EMA designs limit identification of the most influential contextual predictors. Therefore, the present study aims to examine associations between multiple drinking context predictors and both typical and high-risk drinking outcomes to identify the most impactful contextual factors for inclusion in brief EMA assessments.
The study recruited 528 undergraduate students (ages 18-24) who had consumed alcohol in the past 2 weeks from a large public university as part of a randomized controlled trial examining online intervention effectiveness, with participants completing baseline, 1-month, and 3-month follow-up surveys. At each wave, participants reported on their most recent drinking occasion, including consumption quantity, duration, blackouts, passing out, and perceived drunkenness, along with contextual factors: environmental context (location using select-all options), social context (number and proportion of people drinking, relationship types, social interaction), alcohol offers (four types of drink offers assessed dichotomously), and drinking motives (modified from the Drinking Motives Questionnaire into four categories: social, enhancement, coping, and conformity motives rated 0-4). Multilevel modeling was conducted on the full sample with separate models for each context type and outcome, focusing on within-person prediction across 30 final models.
The findings indicated that contextual factors differentially predicted typical and high-risk drinking outcomes. For typical drinking outcomes (drinks consumed, eBAC, perceived drunkenness), drinking away from home (e.g., in a bar, someone else’s home), having more people present, and drinking with friends increased consumption and intoxication, while drinking with family was protective. Alcohol offers emerged as one of the strongest predictor sets, with social and enhancement motives also consistently associated with greater consumption and intoxication. For high-risk outcomes (binge drinking, blacking out, passing out), drinking in multiple locations and social factors showed selective associations across different outcomes (e.g., interacting with others predicted binge drinking and blacking out but not passing out), while alcohol offers again emerged as strong predictors with different offer types showing specific patterns. Social predictors and alcohol offers consistently accounted for the most within-person variance across outcomes (typically 10-15%), with alcohol offers explaining particularly high variance in blackout drinking (45%), while location generally explained the least variance (5-7%).
Takeaway: Social drinking contexts—particularly who you’re with, group size, and receiving unsolicited drink offers—are the most powerful predictors of both heavy drinking and alcohol-related harm, suggesting that effective college drinking interventions should focus on helping students recognize and navigate these high-risk social situations.