Recently findings that marital status is associated with survival in patients with late-stage prostate cancer led to an examination of the generalizability of this association for all cancers. We restricted the investigation to patients with late-stage cancer using population-based data collected from 261,070 patients with late-stage cancer at multiple sites in the United States to determine relations between marital status and survival. After controlling for age, race, and treatment, married patients with cancers of all major primary sites had significantly better survival than single, separated, divorced, or widowed patients. Although single and widowed patients had the poorest prognosis in general, single patients appeared to show the most consistently poor survival across the different types of cancers. Survival differences by marital status were more pronounced in men than in women. This observation raises the possibility that some characteristics associated with being married delay death from cancer. These findings require investigators to ask new questions about the effect of being married and its possible correlates, such as general health status, access to health care, and socioeconomic status. Known correlates of marital status, such as available social support and social isolation also merit attention in relation to these findings.