Recent publications have reported a weak but consistent correlation between diabetes incidence and occupational or accidental exposure to dioxins. As with most work involving environmental xenobiotics, these studies suffered from the analytical problem that the reference populations had some degree of exposure. We have used Haber's Rule to relate the integrated exposure of subjects involved in an industrial exposure to dioxins, reported by Sweeney et al. [Teratog. Carcinog. Mutagen. 17 (1998) 241], to the incremental probability of diabetes incidence. We estimated that background exposure to dioxin-like compounds by the referents contributed <1% of their diabetes risk, suggesting that background exposure to dioxins is not a significant risk factor for individuals who have not been occupationally or accidentally exposed.