Low-level Radiation and Infant Mortality

Abstract

Abstract

Annual data for 1960–70 for each of the 48 contiguous states are analyzed in a variety of ways to determine whether interstate or intertemporal differences in infant mortality were associated with differences in radioactive fallout as measured by 90Sr or 137Cs in pasteurized milk. Typically, the null hypothesis of no effect of fallout on infant mortality cannot be rejected. This conclusion is relatively robust under alternative specifications of time trends, the period of study, the timing of the relation between radioactivity and mortality, and the inclusion or exclusion of other variables. Changes in radiation levels do not explain the slow improvement in infant mortality in the early 1960s or the rapid improvement after 1965.