Epidemiologic designs (case control, cohort, and surveillance) and measures of association are discussed in the context of studies in the Synthetic Vitreous Fibers industry. Current investigations including a case control study for Owens-Corning's Newark, Ohio plant and a Corporate-wide Mortality Surveillance System are described. The importance of internal validity and the need to account for confounding are demonstrated. Estimates of the prevalence of smoking developed for the Newark plant and the U.S. population have been used to adjust a previously reported statistically significant lung cancer Standardized Mortality Ratio (SMR) for Owens-Corning's Newark plant. After adjusting for the confounding effect of cigarette smoking, the Newark SMR based on national mortality rates is reduced to a statistically insignificant 107.7, a level quite similar to that obtained using local mortality rates. While smoking does not account for all of the excess lung cancer mortality in this population compared to U.S. mortality, these investigations make it clear that plant exposures including respirable glass fibers are not responsible for the remaining excess; rather, this excess reflects the effect of some unknown set of social, demographic, or chance factors.