“Characterizing Cohort Loss Before Birth”

Abstract: Answers to many central questions in the social sciences depend upon the assumption that cohort loss before birth is ignorable. Evidence from inferential population studies and small-scale cohort studies increasingly suggests otherwise. Up to 70% of human pregnancies terminate before birth; these losses appear to be non-random. In this research we consider the implications of prenatal cohort loss for a few key demographic questions, including the effects of early-life exposures on later-life health and the effects of child traits on parent outcomes. In so doing, we extend a long history of demographic research on cohort selection to the prenatal period. We conclude with a discussion of new, big data approaches to learn more about how prenatal exposures shape population traits.