A Daily-Level Analysis of Moderators of the Association between Alcohol Expectancies and Alcohol Use among College Student Drinkers

Abstract

Background: Alcohol expectancies, or the perceived likelihood of experiencing certain effects after consuming alcohol, are associated with college student drinking such that heavier drinkers expect a greater likelihood of positive effects. However, less is known as to whether day-to-day within-person deviations in expectancies are associated with drinking that same day and for whom and when these associations may be strongest. Objectives: The aim of this study was to examine daily-level associations of positive and negative alcohol expectancies with alcohol use, and whether associations differed according to demographic characteristics and additional alcohol-related constructs. Methods: College student drinkers (N = 327, 53.8% female) participated in an intensive longitudinal study that captured daily-level data. Alcohol use and expectancy measures were utilized from a baseline session and at the daily-level using Interactive Voice Response (IVR). Results: Results found that on days when participants reported stronger positive and negative expectancies than their average, they were more likely to drink as well as consume more alcohol when drinking. Moderation analyses revealed that positive expectancies were more positively associated with the likelihood of any drinking for women relative to men, and more positively associated with the quantity of alcohol consumption for younger students, students with lower baseline rates of drinking, and students with greater overall positive alcohol expectancies. Conclusions/Importance: The findings demonstrate that alcohol expectancies fluctuate within-person across days and these fluctuations are meaningful in predicting same-day drinking. Interventions that seek to modify expectancies proximal to drinking events may be considered to reduce college student drinking.

Keywords: Alcohol; college; expectancies; longitudinal; moderation.

Figure 1.

Model-predicted likelihood of any drinking according to daily deviations in expectancies and moderators of interest. Moderator variables include sex (1a), and person-mean positive expectancies across the monitoring period (1b).

Figure 2.

Model-predicted count of drinks on drinking days according to daily deviations in expectancies and moderators of interest. Moderator variables include baseline drinking (i.e. typical drinks per week, 2a), age at baseline (2b), person-mean positive expectancies across the monitoring period (2c), and person-mean negative expectancies across the monitoring period (2d).