Restraint and attributions: Evidence of the abstinence violation effect in alcohol consumption

In this cross-sectional test of the occurrence of the abstinence violation effect (AVE), a community sample of 323 social drinkers completed measures of drinking restraint (perceived efficacy for controlling alcohol consumption and behavioral attempts to limit drinking) and causal attributions for drinking-related events. These measures were included as predictors in a multivariate multiple-regression equation in which three aspects of self-reported drinking served as dependent variables: minimum number of drinks per occasion, maximum number of drinks per occasion, and drinking-related problems. The results indicated differential sets of predictors for each of the dependent variables. Both aspects of restraint predicted minimum consumption, and both aspects of restraint and attributions for positive drinking situations predicted maximum consumption. Perceived efficacy for controlling alcohol consumption and attributions for negative drinking situations predicted alcohol-related problems. The consistency between this pattern of results and Marlatt's (1985a) reformulation of the AVE is discussed.