OBJECTIVES: To determine the impact of physical activity (PA) interventions and experience with PA on psychosocial constructs. METHODS: PA, self-efficacy, self-identity, social support, and expected PA frequency were measured pre and post intervention in 39 persons with multiple sclerosis. RESULTS: The MANOVA analysis indicated that self-efficacy and expectation decreased, whereas self-identity and PA improved (P<0.05). Regressions analyses indicated pretest self-identity (beta(2)=0.44) and social support (beta(2)=0.34) were associated with PA. Posttest self-efficacy (beta(2)=0.38) and social support (beta(2)=0.31) were associated with PA (P<0.05). CONCLUSIONS: The interventions and experience with PA caused changes in psychosocial constructs, which subsequently affected the correlation of these constructs with PA.