Disentangling the Within and Between-Person Aspects of Implicit Alcohol Associations on Hazardous Drinking

Indirect measures have been used to assess the strength of automatically activated, alcohol-related associations and their relation to hazardous drinking. However, little is known about the longitudinal contributions of between-person and within-person components of implicit alcohol associations (IAAs) to college students’ hazardous drinking over time. This study examined how within and between-person variability in three measures of IAA (drinking identity, alcohol approach, and alcohol excite) are related to hazardous drinking while controlling for their explicit, self-report counterparts. First and second-year U.S. college students (N = 506; 57% female) completed web-based assessments once every 3 months up to eight assessments, which included IAA measures (Implicit Association Tests), explicit counterparts of the IAAs, and self-reported hazardous drinking (alcohol consumption, problems, and risk of alcohol use disorder). Bayesian generalized multilevel models were used to examine between and within-person associations among IAA, their explicit counterparts, and hazardous drinking. Results showed that between persons, mean levels of all three IAAs were positively associated with hazardous drinking over time. Once their explicit, self-report counterparts were included, however, only drinking identity IAA remained significant. Within persons, increases in drinking identity IAA were associated with increases in subsequent hazardous drinking risk, even after controlling for its explicit counterpart. These results suggest the importance of disentangling and simultaneously investigating between and within-person processes in IAAs. Although the between-person component of IAAs may play a larger role in the prediction of hazardous drinking, examining the within-person component of IAA, at least for drinking identity, also appears to be important.