College students tend to perceive alcohol use as an integral part of college culture, resulting in higher levels of alcohol use and heavy drinking compared to other young adults, and the associated alcohol-related harms have emerged as a significant public health concern. From a behavioral economic reinforcer pathology perspective, alcohol is conceptualized as an overvalued reinforcer, and the Alcohol Purchase Task (APT) measures alcohol demand (reinforcing value) through the price–consumption relationship, with prior research demonstrating that alcohol demand predicts not only concurrent but also future alcohol use. However, adolescents and young adults often inaccurately predict their future substance use behavior; therefore, the present study examines, among college students who engage in heavy drinking, changes in alcohol demand over time, the accuracy of self-projected future demand, and whether projected demand predicts subsequent alcohol consumption.
This study used a two-wave, 3-month longitudinal design with undergraduate heavy drinkers recruited from a large Midwestern university, with a final analytic sample of 42 participants who completed both sessions. Alcohol demand was assessed using a standard Alcohol Purchase Task (APT) at both time points and a projected APT at baseline to estimate demand three months in the future, alongside measures of alcohol use and related problems using the AUDIT and past-month consumption via the Timeline Follow-Back. Data were screened for quality and analyzed using exponential demand modeling, delta change scores, correlations, and regression analyses to examine changes in demand, the accuracy of projected demand, and whether projected demand predicted subsequent alcohol consumption.
Over the 3-month period, participants’ observed alcohol demand indices did not change significantly, and each T1 index was significantly positively correlated with its corresponding T2 index, indicating stability over time. In contrast, when asked at T1 to consider their behavior three months later, participants projected that they would purchase/consume more alcohol and be less sensitive to price. Projected values were significantly positively correlated with both observed T2 values and observed change scores, indicating relative (moderate) accuracy in projections, and mean-level differences between projected and observed values were generally non-significant. However, projected maximum expenditure was significantly higher than observed, indicating overestimation for this index. In addition, in unadjusted analyses, projected consumption at zero price (intensity) and projected total expenditure significantly predicted past-month alcohol consumption at T2; however, these associations were no longer statistically significant after controlling for T1 alcohol consumption or baseline alcohol-related risk.
Takeaway: Among heavy-drinking college students, projected alcohol demand for three months ahead was higher than current demand and was positively associated with subsequent alcohol consumption.