AbstractThe role of stress and related factors in cancer etiology and progression has been a subject of conjecture and hypothesis. We analysed the association of type A/B behavior, which can be viewed as a chronic, repetitive stress response, and cancer mortality, using data from a 22‐year follow‐up of the Western Collaborative Group Study. The unadjusted relative hazard between type A/B behavior and total cancer mortality was 1.53 (p= 0.005), indicating an excess risk of cancer mortality in type A subjects of about 50% over that of type B. However, most or all of the association appeared to be confounded or mediated by other risk factors. Cigarette smoking and age were only weakly correlated with type A/B behavior, yet these two variables were so strongly related to cancer mortality that they accounted for about half of the association of type A/B with cancer mortality. The remaining association was found only for mortality of those cancers which are linked to alcohol consumption, suggesting that this remaining association may operate through a link between type A behavior and alcohol consumption. Thus, an unadjusted association between type A/B behavior and cancer mortality was mostly accounted for by cigarette smoking, age, and possibly alcohol consumpt