Our aim was to determine the predictive value of some psychometric instruments for the development and persistence of musculoskeletal pain. In 452 subjects, pain in the shoulder, neck, and low back during the preceding year was assessed at baseline, one year, and two years' follow-up. Psychological distress, depression, self-efficacy beliefs, subjective work prognosis, disability, and work characteristics were assessed at baseline. The best predictor of future pain was disability. The psychometric measures did not predict changes in pain. The explanatory power of the variables in the multivariate analyses was low. Perceived disability in persons with musculoskeletal symptoms should be acknowledged early. The traditional research paradigm focusing on a few hypothetically relevant variables does not take into account the reflective, dynamic, reciprocal nature of human behavior during the process of recovery from or development of a pain problem.