Background. We have recently demonstrated that individual psychosocial factors are important predictors of the onset of chronic widespread pain. It has been hypothesized that excessive mechanical exposure may also be associated with symptom onset, although this has not been formally examined. We therefore determined the relative contributions of individual psychosocial and work-related mechanical, posture and environment factors in symptom onset. Methods. We conducted a population-based prospective survey and identified 1658 adults aged 18–65 yr who were symptom-free. At baseline, detailed information was obtained on work-related mechanical and environment factors using validated instruments. Individual psychosocial features were also measured. Subjects free of chronic widespread pain at baseline were followed up at 12 and 36 months to identify those reporting the onset of new symptoms. Results. In all, 1445 (91%) returned the questionnaire at 12 months and 978 (89%) at 36 months. Of these, 81 and 92 respectively reported new chronic widespread pain. Symptom onset was predicted by workplace factors {pushing/pulling heavy weights [relative risk (RR) = 1.8, 95% confidence interval (CI) 1.1, 3.0]; repetitive movements of the wrists (RR = 1.8, 95% CI 1.2, 2.7); kneeling (RR = 2.2, 95% CI 1.2, 4.1)} and individual factors [aspects of illness behaviour (RR = 2.9, 95% CI 1.6, 5.3); somatic symptoms (RR = 1.9 95% CI 1.1, 3.3); fatigue (RR = 1.9, 95% CI 1.2, 3.1); baseline pain symptoms (RR = 2.5, 95% CI 1.6, 3.9)]. In multivariate analysis, pushing/pulling heavy weights, repetitive wrist movements, kneeling and other pain at baseline were associated with new-onset chronic widespread pain. However, the strongest predictor was a high score on the illness behaviour scale. Conclusion. This study provides only limited support for the hypothesis that low-level mechanical injury may be a risk factor for developing chronic widespread pain. The onset of chronic widespread pain appears to be multifactorial and is strongly predicted by individual psychosocial factors.