Context Matters for the Socially Anxious: Moderating Role of Drinking Context on Alcohol Outcome Expectancies

Abstract

Background: Moderating effects of alcohol outcome expectancies (AOE) on the social anxiety (SA)-alcohol misuse relationship are mixed. This may be explained by differential relationships between SA and context-specific AOE. Gender may further moderate these associations, as it influences SA, AOE, and drinking behaviors. Objectives: To examine the moderating role of drinking context (i.e. convivial, negative coping, or personal-intimate) and gender on the relationships between SA and three AOE (i.e. tension reduction, sociability, and sexuality). Methods: Participants (n = 436, M_age=19.32, 72% female, 85.8% White) were 218 undergraduates with elevated SA (high SA group) and a gender-matched low SA group (n = 218) drawn from a larger undergraduate sample (N = 1,015). Participants completed three versions of an AOE measure, differing by drinking context considered. Results: A significant SA group x context x gender interaction was found for tension reduction AOE; compared to men, women in the low SA group reported greater tension reduction AOE in negative coping contexts. Significant SA group and context main effects suggest that sociability and sexuality AOE are endorsed more in the high (vs. low) SA group, and in convivial and personal-intimate compared to negative coping contexts. Conclusions/Importance: Tension reduction AOE vary depending on the drinking context, SA, and gender. Assessment of AOE in specific drinking contexts may help to identify which individuals may be at greatest risk for alcohol misuse and help inform treatment of SA-related problem drinking.

Keywords:

• Alcohol
• expectancies
• social anxiety
• context
• gender

Notes

1 We would like to thank the anonymous reviewer for the suggestion to conduct these post-hoc analyses.

2 We also conducted analyses testing SA x context x gender interactions on sociability AOE. Main effects of gender and all interaction effects were non-significant. There were significant main effects of SA group, F (1, 432) = 25.12, p < .001, η_p² = .055, and of context, Wilk’s Lambda = 0.967, F (2, 431) = 7.40, p = .001, η_p² = .033, with the same pattern differences present as when gender was excluded from the model.