Many work-site physical activity interventions use theoretical variables in the design of their programs. Yet, these interventions do not document the degree of change in theoretical variables produced by the intervention. This study examined the construct validity of an intervention designed to affect social cognitive theory variables linked to exercise behavior. Construct validation methods were used to evaluate an instructional intervention composed of four 60-minute sessions delivered across 2 weeks. Increases were found in self-regulation skills, outcome-expectancy values, and self-efficacy for the treatment group. No significant increases were detected for the comparison group on any study variables. Sixty-seven percent of the treatment group was able to maintain exercise behavior across 12 months, whereas the comparison group declined in exercise participation from 68% to 25% across 12 months. The study revealed the intervention effective in producing the intended changes in social cognitive theory constructs. The analysis shows self-regulation-mediated exercise behavior. (Contains 4 figures.)