To systematically evaluate bias in estimating relative risks associated with occupational exposures obtained in the absence of smoking data, the authors compared crude and smoking-adjusted standardized mortality ratios for selected occupations using data from a study of U.S. veterans. Crude and smoking-adjusted SMRs were highly correlated: lung cancer (r = .88), bladder cancer (r = .98), and intestinal cancer (r = .97). Greater differences occurred for lung cancer, which is more strongly related to smoking than bladder or intestinal cancer.