Laboratory

Population Dynamics and Sustainable Well-Being

At a Glance Projects Publications Team

Project

Estimating and Modeling Fertility Indicators

Benjamin-Samuel Schlueter, Jessica Donzowa, Emilio Zagheni, Daniela Perrotta

Detailed Description

This project is based on the fact that timely and accurate information about fertility indicators is not always available for specific subpopulations. The quality of vital statistics varies by country and continent. While the coverage levels of vital registration systems are high in high-income countries, they are lower in low- and middle-income countries. Moreover, even in contexts with high coverage levels, the necessary information to develop an indicator is not always publicly available. In these instances, we must rely on surveys.

Using survey data to estimate indicators for a subpopulation leads to well-known limitations, such as large sampling errors due to small sample sizes and changes or interruptions in data collection. In these contexts, it is valuable to develop statistical models that can estimate or reconstruct fertility indicators while accounting for missing data, small sample sizes, and data quality issues. Bayesian models are especially useful because they allow us to smooth estimates, borrow information between units, incorporate demographic knowledge, and combine surveys.

Below, we provide examples of research lines in this project, including the estimation of male age-specific fertility rates in low- and middle-income countries; the estimation of the proportion of childless women by age, race, ethnicity, and education in the US; and the combination of online surveys and network-reporting methods to estimate fertility indicators in Senegal.

The biggest differences in timing and magnitude between male and female fertility are observed in the Global South, where data on male fertility are not widely available. In this project, we propose a Bayesian parametric model to estimate and reconstruct male fertility rates for countries with no civil registration and vital statistics systems. This research will contribute to a more comprehensive understanding of male fertility trends and provide essential inputs for modeling kinship structures and orphanhood, and for generating indirect mortality estimates. In future work, we plan to estimate male age-specific fertility rates by education level in low- and middle-income countries.

Childlessness across the life course is an increasingly important demographic phenomenon. We developed a Bayesian parametric model to estimate the proportion of women who are childless by age, race and ethnicity, and education for US cohorts born in 1950-1999. We show that there have been substantial changes to childbearing trajectories in the United States, with an increase in the share of women who are childless at most ages. In future work, we plan to develop the model further in order to estimate cohort trends in parity-specific proportions of women by age, race, ethnicity, and education in the US.

A third line of research integrates social media-based recruitment with a network reporting approach to assess the potential of online surveys for generating reliable population-level indicators in low- and middle-income contexts. In this approach, respondents were asked to share information about people in their social networks. The Senegal Demographic Survey, which we conducted in October 2024 via Facebook ads recruitment, demonstrates how online surveys can help estimate fertility indicators in Sub-Saharan Africa, and thus contributes to the growing literature on digital data collection in low- and middle-income countries.

Prediction of mean age at childbearing for men in 2017 based on data from Facebook’s Advertising Platform

Prediction of Male mean age at childbearing in 2017 for countries without UN ground truth data. © MPIDR

Publications

Donzowa, J.; Perrotta, D.; Feehan, D.; Zagheni, E.:

Fertility measurement in data-constrained settings: possibilities and limitations of a Facebook-recruited network reporting survey in Senegal

MPIDR Working Paper WP-2026-030. (2026)    

Schlueter, B.-S.; Bruno, S.; Alexander, M. J.:

A Bayesian parametric model to estimate and reconstruct male age-specific fertility rates

MPIDR Working Paper WP-2026-006. (2026)    

Schlueter, B.-S.; Root, L.; Alexander, M. J.:

Trends and patterns in childlessness by age, race, ethnicity, and education in the United States: a research note

MPIDR Working Paper WP-2026-005. (2026)    

Donzowa, J.; Kühne, S.; Zindel, Z.:

From clicks to quality: assessing advertisement design’s impact on social media survey response quality

Methods, Data, Analyses 19:1, 95–136. (2025)    

Donzowa, J.; Perrotta, D.; Zagheni, E.:

Assessing self-selection biases in Facebook-recruited online surveys: evidence from the COVID-19 Health Behavior Survey

PLOS One 20:7, e0326884–e0326884. (2025)    

Rampazzo, F.; Zagheni, E.; Weber, I. G.; Testa, M.R.; Billari, F. C.:

Mater certa est, pater numquam: what can Facebook advertising data tell us about male fertility rates?

In: Proceedings of the 12th International AAAI Conference on Web and Social Media (ICWSM 2018): 25-28 June 2018, Stanford, California, 672–675. Palo Alto, CA: AAAI Press. (2018)