Previously sedentary women in the southeast United States (M age = 35.9 years, SD = 10.1), whose primary reason for beginning an exercise program was either health improvement (N = 48) or appearance improvement (N = 42), were assessed to estimate relationships between perceived progress on these factors, self-motivation, and exercise program attendance. Non-significant bivariate correlations were found between attendance and changes in measures of perceived health (r = -.24) and appearance (r = -.02), over the 16-week study. Self-Motivation Inventory scores were supported as a significant correlate of exercise attendance (rs = .30 and .36), and when entered into linear multiple regression analyses, along with perceived health and appearance improvement, explained variances in attendance increased (R2 = .11 and R2 = .14, respectively). Findings supported trait theory as a valid but incomplete model, and failed to support propositions embedded within goal setting and self-efficacy theories. The dynamic relationship between personal, process, and outcome factors, related to exercise adherence, is discussed. The need for further research to contribute to more comprehensive exercise behavior models, capable of giving effective direction to treatment design, is suggested. Limitations and the need for replication and extension are also discussed.