Work-related musculoskeletal disorders account for the largest single category of lost-time occupational injury or disease episodes in industrialized countries. In this study we analyzed the impact of the psychosocial work environment, conceptualized by the demand-control-support model, on the incidence of low back and neck injury in a cohort of 1221 public transit operators followed for 7 years and 6 months. The two main exposure variables were "job strain" (mismatch of high psychological demands and low decision latitude) and "iso-strain" (job strain plus exposure to low social support at work). Analyses controlled for demographic factors, physical workload, and pain at baseline. For low back injuries, increased hazard rates were found for job strain and iso-strain based on tertiles, with hazard ratios (HR) of 1.30 (95% CI=0.96-1.75) and 1.41 (95% CI=0.98-2.01), respectively. Job strain and iso-strain based on median split or analyzed as continuous variables were not associated with low back injuries. For neck injuries, job strain and iso-strain based on median split showed HRs of 1.27 (95% CI=0.99-1.63) and 1.33 (95% CI=1.01-1.77), respectively. Job strain and iso-strain based on tertiles had HRs of 1.52 (95% CI=1.13-2.05) and 1.73 (95% CI=1.21-2.45), respectively. When analyzed as continuous variables, a one-point increase on the job strain and iso-strain scales led to an 8% (95% CI=0.98-1.19) and 14% (95% CI=1.02-1.27) increased hazard of neck injuries, respectively. This study shows the importance of the psychosocial work environment in the etiology of musculoskeletal injuries among transit operators. Since reviews have shown that psychosocial workplace conditions in this occupational group can be changed, these findings indicate a potential for prevention at the workplace.