Modeling Age Patterns of Childlessness: A Bayesian Parametric Approach

Benjamin-Samuel Schlüter, University of Toronto
Monica Alexander, University of Toronto

We present a Bayesian hierarchical parametric model to estimate age-specific proportions of women who are childless. Central to the model is the Janoschek growth curve model, which is parameterized by three components that relate to the eventual level of childlessness, the rate of decline of childlessness over age, and the age at which childlessness begins to stablize. We incorporate this growth curve model into a Bayesian framework that allows for information sharing across subgroups and over time. We also account for known biases in existing survey data. We apply the model to estimate childlessness over age in the United States over the period 1976 to 2022, and by race/ethnicity.

Keywords: Bayesian methods , Fertility, Population projections, forecasts, and estimations, Small area estimation

See paper.