February 19, 2025 | News | Recommended Reading
New Approach Predicts Individual Fertility Behavior from Aggregated Data
In their paper published in November, Daniel Ciganda and Nicolas Todd present a new approach to identifying individual fertility behavior from aggregate data. This method offers an alternative to traditional macro-level models and provides a better understanding of the drivers that determine fertility patterns.
© istockphoto.com / powerofforever
In the study, individual-level reproductive parameters such as age at marriage, age-specific fecundability and duration of postpartum amenorrhoea were reliably predicted from population-level fertility rates. These results suggest that individual-level fertility dynamics can be reliably predicted even in the absence of individual-level data, at least in natural fertility settings. 'Using our approach, we were able to accurately capture the reproductive patterns of three historical populations with natural fertility and identify the key individual-level factors that shape population-level fertility fluctuations,' explains Daniel Ciganda. He is leader of the Max Planck Partner Group Simulation of Sociodemographic Systems in Montevideo, Uruguay hosted by the Statistics Institute of the University of the Republic (UDELAR).
Estimated versus observed distributions of age at marriage: Purple bars represent the observed age at marriage in 17th-18th century France, red shaded areas represent the age-distribution estimated by the model. © MPIDR
The proposed approach provides an interpretable framework for analyzing fertility trends, particularly in populations where individual-level data are limited. It fills a critical gap in demographic research by linking aggregate data with individual-level behavioral and biological mechanisms. Ciganda and Todd's new approach is particularly valuable for understanding historical and current fertility dynamics.
'Although our study is limited to the simplest case of natural fertility, we are currently working on an extension that incorporates the more complex set of behavioural and biological factors that shape fertility dynamics in contemporary societies' explains Ciganda.
'Because it can provide a deeper understanding of fertility behavior without relying on extensive individual-level data, our approach provides advantages over more traditional methods, both for modeling and forecasting of fertility dynamics ' explains the researcher.
Original Publication
Ciganda, D.; Todd, N.:
Royal Society Open Science 11:11, 1–14. (2024)
Keywords
Agent-based modeling, fertility, reproductive behavior, demography, approximate Bayesian computation, age-specific fertility rates, likelihood-free inference, simulation