This study examined the effects of self-monitoring behavior and health locus of control on improvement in a work hardening program. The subjects included 22 male and 18 female outpatients in a hospital-based rehabilitation program. Subjects were classified as having an internal or external health locus of control, and were randomly assigned to either a self-monitoring or a non-self-monitoring group. Improvement was assessed via objective performance data and self-ratings of perceived improvement. The results indicated that individuals identified as having an internal health locus of control did not show greater gains in physical functioning or perceived improvement relative to externally oriented individuals. Additionally, those subjects participating in self-monitoring activities were no different from non-self-monitoring subjects in terms of improvement in exercise activities or perceived improvement. The results also indicated no interaction between health locus of control and the presence or absence of self-monitoring. It was suggested that other factors such as workman's compensation, pain patient characteristics, low self-concept, and severe stress may have proved more powerful influences on patient improvement than internal health locus of control or self-monitoring. It was also suggested that rehabilitation programs might benefit from creating structured environments in which patients receive frequent staff feedback and reinforcement for improvement. Monitoring small, discrete, easily attainable goals might prove more effective than monitoring mood, pain, etc. In addition, teaching specific internal health locus of control behaviors to patients may help them improve their self-concepts and progress. Further research is needed to explore the roles that pain patient personality characteristics, self-concept, and stress play in the progress of patients in a work hardening program.