ABSTRACT Using an expanded Social Cognitive Theory (SCT) model, we hypothesized that self-efficacy, outcome expectations, and exercise self-definition would predict exercise adoption. This secondary analysis examined data from a prospective single-group study of low-income women who received a physician screen and referral to a community-based, free exercise program. The sample included 190 older, low-income women with a mean age of 64 years, the majority of whom were African American (66%) and had at least one cardiovascular risk factor (92%). Baseline values of self-efficacy, outcome expectations, and exercise self-definition were measured using instruments developed for the study. Exercise adoption was defined as the number of exercise sessions completed over 8 weeks. Our hypothesis was tested using hierarchical multiple regression. The mean number of exercise sessions completed over the 8-week period was 5.7 out of a recommended 24. Value of Exercise scores, a subscale of the Exercise Self-Definition scale, predicted exercise adoption. Self-efficacy and outcome expectations were not predictive. The significance of Value of Exercise scores reinforces the importance of expanding SCT with additional variables such as exercise self-definition. Future work should emphasize the social and environmental factors that form an important part of SCT.