Some exhaustible Poisson process models of divorce by marriage cohort†

Publisher Website
Google Scholar
Abstract
A one‐parameter exhaustible Poisson process model is formulated to represent the cumulative divorce trajectory of marriage cohorts. On the basis of recently published data of nine‐year cumulative records of all one‐year United States marriage cohorts, 1949–1958, it is found that the one‐parameter model does not provide an empirically adequate representation of cohort divorce trajectories. Therefore, two modifications of the basic model are explored. First, it is assumed that the longer the couples remain married the smaller is the probability of their becoming divorced (a cumulative inertia modification). Second, it is assumed that the cohort can be divided into two groups one of which is subject to risk of divorce while the other is not (a mover‐stayer modification). It is concluded that the latter model provides a better approximation to the empirically observed divorce trajectories. Given availability of disaggregated observations of divorces for marriage cohorts, the model could be used as a mathematical language for the construction of a theory of divorce differentials and changes in the divorce condition over time.