Religious differentials in infant and child mortality in Holland, 1855-1912.

At least three kinds of hypothesis may be invoked to interpret religious differentials in mortality. They are (i) hypotheses that refer to characteristics, (ii) those that refer to lifestyle, and (iii) those that refer to the social isolation of minorities. This paper tests all three kinds of hypothesis using data on urban child mortality from the Hague just before and during the demographic transition. A hazard analysis suggests that economic and demographic characteristics do not account for much of the variation by religion. An analysis of seasonal mortality suggests that some of the variation may be explained by differences in lifestyle. The third kind of hypothesis is presented here for the first time. We suggest that the social isolation of small religious groups lowered their exposure to certain kinds of infectious disease. We use a simulation study to show that this hypothesis could account for part of the variation.