Statistical analyses of 1,475 histologically confirmed cases of malignant mesothelioma ascertained through the Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute for the years 1973-1984 showed age at diagnosis, sex, stage of disease, type of treatment, and geographic area of residence to be important predictors of patient survival, although type of treatment may be confounded with prognostic factors (patients selected for surgical treatment tended to have better performance status than other patients). Women below the age of 50 had an unusually long survival, even after adjustment for the effects of other variables in the model. A relatively large proportion of female cases had site of disease designated as peritoneum, but site was not a significant prognostic factor. These results suggest that age, gender and stage of disease should be carefully considered in designing and analyzing clinical trials for persons with mesothelioma. Survival was shorter in the 4 SEER registries which had shipbuilding as a major industry than in the others with less potential asbestos exposure, offering weak support for the hypothesis that asbestos-exposed cases of mesothelioma have worse survival experience than other cases.