Modeling Abortion as a Process, with an Application to French Survey Data

Clementine Rossier, Institut National d'Études Démographiques (INED)
Nathalie Bajos, Institut National de la Santé et de la Recherche Médicale (INSERM)
François Michelot, Institut National de la Santé et de la Recherche Médicale (INSERM)

We model the proximate determinants of induced abortion as a nested sequence of events and behaviors, unlike cross-sectional models that view abortion decisions as being made only in the moment. Our approach allows the investigation of the effect of combinations of background variables on the decision to have an abortion, which may take different directions as the process unfolds, and helps to explain phenomena such as the inverted-U shape of abortion with age. We illustrate the model using data from a reproductive health survey conducted in France in 2000.