Designs for studies of personal exposure to air pollution and the impact of measurement error

This article discusses the impact of measurement error on various designs for the study of the health effects on humans of exposure to air pollution. It discusses time-series studies in which total deaths (or other events) are related to day-to-day variations in air pollution level, and relates the power of these studies to the power of more complex studies in which geographical variation and data on individual exposures are included in time-series analysis, and in which individual outcomes are included as well. Much emphasis is on the recent analysis of Sheppard (this issue) as a source of information concerning variations in grouped and individual air pollution.