As a result of political campaigns by women’s advocates during the 1970’s, offender behavior programs for perpetrators of domestic violence began emerging. However, while most violent offender programs focus on re-education and skills training, many of the domestic violence offender programs focus on information presentation. Moreover, evaluations of these programs have focused too narrowly on the one outcome measure of recidivism. Although recidivism illustrates the impact of a program on offending behavior, it does not reveal how the program elicited change. The authors argue that evaluations of domestic violence offender programs should adopt a more holistic approach that incorporates investigation of the psychological characteristics of offenders and treatment characteristics that work, for whom, and under what conditions. Toward this holistic approach, the authors outline their interpretation of Rossi, Freeman, and Lipsey’s comprehensive evaluation framework, which was originally published in 1979 and has since become a generic guideline for evaluators. Under this guideline the process of evaluation would benefit from assessing the characteristics of offenders, their potential heterogeneity, and their resulting response to treatment. Specifically, the relationship between pretreatment motivation to change, program attendance, and behavioral change requires more inquiry. Program implementation and integrity variables also require further inquiry concerning how they relate to the mediating variable of the therapeutic environment. By incorporating an analysis of individual change and aspects of the therapeutic environment associated with change, understanding of the mechanisms responsible for both positive and negative behavioral outcomes will be enhanced and may lead to more effective program design in the future. Only by incorporating holistic approaches to program evaluation will evaluators approach an understanding of the process of successful rehabilitation. Tables, references