The influence of drinking consequences on alcohol expectancy likelihoods and valences: an item-level multi-level approach.

Abstract

Objective: Alcohol expectancy theory proposes that beliefs about drinking motivate or deter drinking. Although expectancies influence drinking, less is known about how the consequences of drinking influence expectancies. We modeled a feedback conceptualization of how the experience of specific consequences influenced people's beliefs about how likely a consequence will occur (i.e., likelihoods) and how positive the consequence will be (i.e., valences).

Method: We re-analyzed cross-sectional data from college drinkers (n = 504), using Bayesian cross-classified multilevel ordinal regressions to estimate associations between consequences, likelihoods, and valences. We performed a preregistered replication in new data (n = 362).

Results: Participants had higher likelihoods (95% CI_{Study 1} = [2.06, 2.43], 95% CI_{Study 2} = [1.75, 2.12]) and valences (95% CI_{Study 1} = [0.28, 0.52], 95% CI_{Study 2} = [0.33, 0.60]) when they had experienced consequences more often, but these associations leveled off at higher consequence frequencies. Participants also believed consequences to be more likely when they viewed them as more positive, and vice versa, and again, these associations leveled off at higher levels of the predictor. Critically, the strength of these associations varied across both people and consequences.

Conclusions: Experiencing specific consequences more often was related to people judging them to be more likely and more positive in the future, aligning with alcohol expectancy theory. This may lead to experiencing negative consequences repeatedly because people are not being demotivated from drinking in the same fashion. Given the person and consequence level variability, clinicians should consider an individualized approach when targeting drinking consequences.