We tested the hypothesis that job strain (the combination of high psychological job demands and low work control) is positively associated with smoking prevalence and intensity in a study group of 389 males employed in a chemical plant, using a self-administered questionnaire. In a logistic regression analysis which controlled for a number of sociodemographic factors, job strain was not found to be associated with smoking cessation. However, among smokers, those in higher-strain jobs smoked more heavily than those in lower-strain positions (OR 1.70, 95% CI = 1.10, 2.61) and were more likely to have increased the amount they smoke (OR 3.72, 95% CI = 1.92, 7.17).