Objective: The purpose of this study was to (a) investigate differences in demographic and psychological variables between treatment completers and dropouts among abusive men entering a court-mandated treatment program and (b) evaluate the goodness of fit of a logistic regression model for predicting attrition developed on a 16-week program by applying that model to the 26-week program. Method: The study employed a secondary analysis of 1,702 men, 850 completers and 852 dropouts, randomly selected from a larger pool of 3,595 men in court-ordered treatment. Results: Analysis indicated that few of the demographic and psychological variables differentiated between completers and dropouts. A logistic regression model for the 26-week program was developed that correctly predicted treatment completion for 60% of the sample and that employed some of the same predictor variables as the 16-week model. Conclusion: Implications for predicting attrition using logistic regression models for enhancing retention rates were explored and discussed. (Contains 5 tables.)