Epidemiological issues related to the study of pneumoconiosis among underground coal miners were discussed. Two major studies were the Pneumoconiosis Field Research (PFR) and the National Study of Coal Workers' Pneumoconiosis (NSCWP). Neither study was considered representative of all miners. The study designs discussed included cross sectional, longitudinal, and case/control studies. Both the PFR and the NSCWP began as cross sectional studies. The healthy worker effect and the lack of information concerning disease development were viewed as disadvantages of cross sectional studies. In contrast, the longitudinal study was described as incorporating temporal factors and often providing better exposure data. However, limitations of the longitudinal study included the healthy worker effect and complicated analyses. Case/control studies were considered optimal for the study of rare diseases. Important selection issues were reviewed, such as the inclusion of exminers and nonparticipants in analysis. In estimating the exposure response relationship, cumulative exposure, a function of both dust concentration and exposure duration, was regarded as the most suitable exposure metric for Coal Workers' Pneumoconiosis (CWP). The exposure response models applicable to CWP included the threshold model, the zero risk at zero exposure model, and the nonzero risk at zero exposure model. Issues related to the choice of outcome variable were also mentioned. Exposure estimation was viewed as a critical, albeit weak, aspect of epidemiological study. The widespread use of questionnaires and radiographs presented problems concerning data reliability and validity. The authors conclude that many methodological problems have been overcome in the epidemiological investigations of CWP among coal miners.