ABSTRACT
In the setting where the researcher wishes to relate disease outcomes to pollution, count measurements arise naturally because often the disease is relatively rare and because the time scale or geographical scale on which pollution measurements fluctuate are necessarily small. For example, in a study of the relationship between temperature and the incidence of sudden infant death syndrome (SIDS) Campbell worked with daily measurements of temperature and SIDS counts. If the disease is even rarer (e.g., polio) the time scale that is suitable might be one month. Aggregation over longer time periods in both of these examples would obliterate or at least blur the link between the putative risk factor and the disease outcome.
