We examined the relation of stressful life events to the risk of colorectal cancer in a population-based case-control study conducted in Stockholm from 1986 to 1988. Completed questionnaires were available for 569 incident cases ascertained through hospital records and the regional cancer registry. Controls were randomly selected from a population register and were frequency matched to the age and sex distribution of the cases; 510 controls completed the study questionnaire. A history of serious work-related problems in the prior 10 years was strongly associated with the occurrence of colorectal cancer. Compared with those with no such problems, the age- and sex-adjusted odds ratio (OR) was 5.5 [95% confidence interval (CI) = 2.3-23.5]. Change of residence was also associated with increased risk: relative to those who had not moved, the OR was 2.8 (95% CI = 1.1-7.1). Similar analyses revealed that increased odds of colorectal cancer were associated with death of a spouse: OR = 1.5 (95% CI = 1.0-2.3). After multivariate adjustment for dietary risk factors and physical activity, these ORs were not substantially altered. Because most of the events were relatively uncommon in this population, the estimates of effect were not precise.