Understanding the dynamics of premature termination from treatment would be useful for practitioners working with clients requiring prolonged treatment. Compliance in weight management was examined in three studies in order to explore the broader issue of compliance. In study 1, a standard 10-week cognitive behavioral weight management program was compared to three specific interventions to increase motivation: monetary contingency contracting (MC), cognitive modification of expectancies (CM), and a combination of both. Subjects were 52 women (aged 20-63) who signed up for the groups according to time preference. The results showed that, as predicted, these interventions decreased dropout and improved weight outcome; differences between groups were not significant. In study 2, motivational variables from expectancy theory were combined with demographic and psychological variables in multiple regression analyses to predict attendance and outcome. Subjects were 27 women (aged 24-69), attending an 8-week weight control class, and the 52 subjects from study 1. One combination of motivational variables, the force to stay and the force to leave, was found to be significantly predictive of attendance. Attendance was highly correlated with outcome, but no combination of variables was predictive of outcome at post-test. In study 3, t-tests were used to analyze differences between a no-measure control group (N=10) and the prediction-only sample (N=27) on attendance and weight outcome. No significant differences were found. Eighteen references are listed. (MCF)