ABSTRACT
In section II we presented an extensive analysis of cross-sectional data for more than one hundred SMSAs in order to explore the association between measurements of air pollution and mortality rates. Implicit in the cross-sectional analysis is the assumption that factors causing increases in mortality, which are not constant across observations, are either accounted for explicitly in the analysis or are uncorrelated with the measured air pollutants. Consequently, to control for factors that cause increases in mortality, we added a host of explanatory variables characterizing demography, home-heating characteristics, climate, and occupation mix, and we found that the results were not inconsistent with the hypothesis that air pollution increases the mortality rate.
