A confirmatory analysis of the hierarchical structure of positive and negative dose-related alcohol expectancies in alcoholics and the associations with family history of alcoholism.

OBJECTIVE The usefulness of measuring four types of alcohol-outcome expectancies in alcoholics was investigated. METHOD The investigation was conducted in three steps. First, a measurement model previously fitted in a general population sample was fitted in the present sample of alcoholics, using confirmative factor analysis. In the hierarchical model tested, four types of expectancies are represented as four second-order factors: positive and negative expectancies for a low and for a high dose of alcohol. The model was compared with competing models. Second, a common measurement model was tested for subgroups of alcoholics categorized by gender and family history of alcoholism. Third, using path analysis, the associations were investigated for the four types of expectancies with variables of potential relevance for treatment. A sample of 451 male and female clinically referred alcoholics volunteered to complete a series of questionnaires, including the expectancy questionnaire. RESULTS Following minor modifications, the measurement model was found to fit adequately in the present sample of alcoholics and across the subsamples. Family history of alcoholism was positively associated with positive expectancies, especially for a high dose of alcohol. This association was mediated by cluster-B personality disorders. CONCLUSIONS Even though the expectancy questionnaire used here should be refined in several respects, the results demonstrate the usefulness of measuring four types of expectancies in alcoholics.