Methods

Abstract Reproductive epidemiology is full of methodological conundrums. Challenges arise from the compressed timescale of events during pregnancy, the opportunity for multiple observations per couple, and the selective effect of past pregnancy outcomes on a couple’s decision to become pregnant again. Some issues are discussed in earlier chapters in the context of specific reproductive endpoints. The present chapter summarizes many of those issues, expands on a few, and introduces some more. Topics here include the use of fetuses at risk for postnatal outcomes, the study of small-for-gestational-age as a surrogate for fetal growth restriction, immortal time bias with time-sensitive exposures, live-birth bias, collider stratification bias (especially in the context of birthweight and gestational age, where the bias is strong), and sibling analysis.