Self-efficacy and alcohol relapse: concurrent validity of confidence measures, self-other discrepancies, and prediction of treatment outcome.

Objective: Self-efficacy has been shown to predict relapse following treatment for alcohol dependence. Most studies use comprehensive multi-item scales to assess clients' confidence. The development and validation of simple measures may encourage both clinicians and researchers to assess self-efficacy more frequently over the course of treatment. However, the validity of both comprehensive and single-item measures is likely to be threatened by deliberate impression management and self-deception, respectively.

Method: One hundred and forty-two alcohol-dependent inpatients completed a shortened unidimensional version of the Drug Taking Confidence Questionnaire and a brief questionnaire on background variables and alcohol use. Additionally, clients' confidence and beliefs about the success of others were assessed using various single-item rating scales. Treatment outcome was evaluated 12 weeks following discharge.

Results: Correlations between confidence measures ranged from r=.21 to r=.56. Abstainers (n=54) differed from relapsers (n=88) with respect to age, marital status, abstinence self-efficacy, and abstinence other-efficacy. Although self-efficacy was not related to treatment outcome, clients' beliefs about the success of others predicted posttreatment drinking behavior.

Conclusions: The present findings suggest that other-efficacy beliefs may reflect an individual's true expectations more accurately than explicit measures of self-efficacy. The predictive validity of self-efficacy measures is likely to be limited because of a positive response bias.