Reducing alcohol consumption during pre-drinking sessions: testing an integrated behaviour-change model

Abstract

Objective: Pre-drinking, the practice of consuming alcohol prior to attending a subsequent event, increases the risk of alcohol-related harm, and is common in undergraduate student populations. The current study tested an integrated behaviour change model to identify the motivational, social-cognitive, and implicit predictors of pre-drinking.

Design: University students (N = 289) completed an online questionnaire comprising measures of motivational and social-cognitive constructs related to reducing pre-drinking alcohol consumption and past behaviour, and an implicit association test for drinking identity. Participants reported their pre-drinking alcohol consumption at follow-up, 4 weeks from baseline.

Main Outcome Measures: Self-reported pre-drinking alcohol consumption.

Results: A variance-based structural equation model revealed that few model hypotheses were supported. Although the effects of past behaviour, perceived behavioural control, and implicit drinking identity, on follow-up pre-drinking alcohol consumption were statistically significant, the effect of intention was not.

Conclusions: Current findings indicate pre-drinking alcohol consumption is associated with past behaviour, perceived behavioural control and implicit drinking identity, and no intentions to reduce pre-drinking alcohol consumption. The finding raise questions over the validity of applying the integrated model in this context. Interventions should consider these factors and attempt to facilitate the formation of intentions that lead to subsequent behaviour.

Keywords:

• Alcohol
• dual-systems model
• pre-drinking
• self-determination theory
• theory of planned behaviour

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Data on the number of participants whose IAT data were excluded due to not meeting the screening criteria were not available due to a software error.

2 A ‘standard drink’ differs between countries. In Australia, a standard drink is a beverage that contains 10 g of ethanol, compared to 14 g in the US, and 8 g in the United Kingdom (Furtwaengler & De Visser, 2013).

3 Relationships between control and model variables are available in the online supplementary materials.

4 We tested whether the model paths and pattern of relationships hypothesised in the integrated behaviour change model differed by gender. Using Satterthwaite Approximation and pooled standard error approaches to compare the path coefficients from each model (Kock, 2014), we found no statistically significant differences.

5 We tested a model excluding the 61 participants who reported consuming alcohol less than once a month (n = 228). Results revealed similar patterns of effects to the overall sample, with slightly larger Beta values associated with the PBC – PDAC (β = −0.26, p < 0.001, f² = 0.06) and D – PDAC (β = 0.16, p = 0.009, f² = 0.02) paths. The effect of controlled motivation on attitude was also statistically significant (β = 0.11, p = 0.049, f² = 0.04), when it was not significant in the overall sample.We also tested a model excluding the 52 participants who did not report consuming alcohol during pre-drinking sessions at baseline (n = 237). Results revealed similar patterns of effects to the overall sample, with the following differences: the controlled motivation – intention path was statistically non-significant (β = 0.03, p = 0.301 f² = 0.01); the subjective norm to intention path was statistically significant (β = 0.11, p = 0.037, f² = 0.05); autonomous motivation to PBC was statistically non-significant (β = 0.08, p = 0.098, f² = 0.01); and the controlled motivation – attitude was statistically significant (β = 0.14, p = 0.017, f² = 0.05) in this sample.

6 The zero-order correlation between intention and behaviour was also non-significant (r < 0.01, p = 0.956).

7 Fishbein and Ajzen (2011) refer to these constructs as capacity and autonomy, respectively.