Many developments in the design and analysis of environmental epidemiology have been made in air pollution studies. In the analysis of the short-term effects of particulate matter on daily mortality, Poisson regression models with flexible smoothing methods have been developed for the analysis of time-series data. Another option for such studies is the use of case–crossover designs, and there have been extensive discussions on the selection of control periods. In the Study on Respiratory Disease and Automobile Exhaust project conducted by the Japanese Ministry of the Environment, we adopted a new 2-stage case–control design that is efficient when both exposure and disease are rare. Based on our experience in conducting air pollution epidemiologic studies, we review 2-stage case–control designs, case–crossover designs, generalized linear models, generalized additive models, and generalized estimating equations, all of which are useful approaches in environmental epidemiology.