The effects of marital status on the diagnosis, treatment, and survival of patients with cancer were examined in population-based data on 27,779 cancer cases. Unmarried persons with cancer had decreased overall survival (relative hazard, 1.23; 95% confidence limits, 1.19 to 1.28). We identified three complementary explanations for the poorer survival of the unmarried persons. First, unmarried persons were more likely to be diagnosed at a regional or distant stage (odds ratio, 1.19; 95% confidence limits, 1.12 to 1.25). After adjustment for stage, unmarried persons were more likely to be untreated for cancer (odds ratio, 1.43; 95% confidence limits, 1.31 to 1.55). Finally, after adjustment for stage and treatment, unmarried persons still had poorer survival. Previous studies have demonstrated that unmarried persons have decreased overall mortality. For cancer, our results suggest that the favorable consequence of being married on overall survival is secondary to the beneficial effects at several steps in the diagnosis, choice of treatment, and response to treatment.