As a test of the "job strain" (job demands-control) model, 297 healthy men aged 30-60 were recruited at eight New York City worksites. The association among job demands and control, social support, and psychological outcomes was tested using both ANCOVA and moderated multiple regression, controlling for demographic variables. The job strain model was supported by various psychological outcome measures, with workers in "active" jobs reporting the highest level of Type A behavior, job involvement, and positive attributional style, workers in "low-strain" jobs reporting the lowest job dissatisfaction and trait anxiety, workers in "passive" jobs reporting the most external locus of control and trait anxiety, and workers in "high-strain" jobs reporting the highest job dissatisfaction. Low social support was associated with greater symptomatology, and a significant three-way interaction (demands x control x support) for job dissatisfaction was observed. While selection of subjects into jobs may partially explain these findings, the results support the hypothesis that working conditions influence psychological attributes and distress.