Social workers are constantly seeking useful data regarding methods of behavior change. Changes in self-concept can mediate changes in overt behavior. Self-concept has been reawakened as a focus of empirical research. Several literatures (self-perception theory, cognitive dissonance theory, and the availability heuristic literature) suggest ways of changing the self-concept. This article reviews existing literature and discusses the guidelines for changing self-concept that derive from empirical research in the previously mentioned literatures. The manner in which these guidelines explain the success of particular intervention techniques is delineated. Suggestions for new social work practice interventions are proffered, as are examples from the authors'practice experience.Clinical social workers are confronted by a diverse variety of approaches to the task of assisting in the process of changing clients' behavior. Each school of therapy has developed a set of interventions to add to the clinician's armamentarium of techniques. The social work clinician is faced with the responsibility of selecting among available techniques, those that will be most likely to achieve success. Sometimes evidence of the efficacy of a particular technique or program of intervention (e.g., assertiveness training) is available. When outcome data exist, the process of selecting a strategy of intervention is made easier. However, even when the social worker employs a validated strategy, an explanation as to why the proven techniques actually work is often not available. Without knowing why a procedure works, the social worker will have difficulty extrapolating to unique situations.Fortunately, empirically validated generalizations regarding human behavior are emerging. These generalizations, empirically supported axioms