July 22, 2025 | News | New Publication

Predicting Migration Flows by Gender Is Difficult with Current Data and Models

Findings reveal significant differences in the predictive performance of gravity-type models based on the gender composition of migration flows

Researchers at the Max Planck Institute for Demographic Research (MPIDR) examined the limitations of theoretical and methodological approaches to understanding gender-specific migration trends. The study's findings reveal that the predictive performance of gravity-type models varies significantly based on the gender composition of migration flows.

Emigrating in hopes of a better life. © istockphoto.com/Naeblys

Published in the high-impact journal Proceedings of the National Academy of Sciences (PNAS), the study evaluates the impact of limitations in methods and migration data on the accuracy of commonly used migration prediction models. “Theory-based methods for migration prediction are limited because migration theories often ignore gender. Meanwhile, data-based methods struggle due to the limited availability of gender-disaggregated migration data,” says Athina Anastasiadou, an MPIDR doctoral student and the study's lead author. Anastasiadou and her colleagues also explore how the gender composition of migration changes depending on the origin and destination of migrants, as well as how migration patterns differ between genders in a global context.

Through an analysis of gender-disaggregated migration flows, the study compares the performance of deterministic methods and probabilistic gravity-type models in predicting migrant flows with varying gender compositions. “We use one of the most comprehensive datasets on international migration, broken down by gender, to compare the accuracy of the results produced by different models based on the predominant gender in migration corridors”, says Anastasiadou.

Why are current prediction methods inaccurate?

The findings reveal significant differences in the predictive performance of gravity-type models based on the gender composition of migration flows. Anastasiadou adds: “Our findings show that widely used gravity-type models yield less accurate results for corridors with a high female gender composition. Based on our method comparison and case study findings, we explain why current prediction methods are inaccurate, citing a lack of sound theoretical concepts and gender-specific data.”

This research underscores the urgent need to reassess migration theories and methods through the lens of gender biases, paving the way for more inclusive and accurate predictions.

“Even though we believe we are being neutral and objective, our methods, theories, and views on migration have been shaped by common narratives and assumptions over the past decades. These narratives center on male labor migrants from the Global South to the Global North. This limits our ability to conceptualize and statistically capture other forms of migration and groups of migrants. Gender biases propagate through all layers of our work, and we must acknowledge this and find ways to overcome these shortcomings,” says Athina Anastasiadou.

Original Publication

Anastasiadou, A., Zagheni, E., de Valk, H.: Including the gender dimension of migration is essential to avoid systematic bias in migration predictions. PNAS (2025). DOI: 10.1073/pnas.2500874122

Authors and Affiliations

Athina Anastasiadou, Max Planck Institute for Demographic Research, Rostock, Netherlands Interdisciplinary Demographic Institute (NIDI) - KNAW/University of Groningen

Emilio Zagheni, Max Planck Institute for Demographic Research, Rostock

Helga A.G. de Valk, Netherlands Interdisciplinary Demographic Institute (NIDI) - KNAW/University of Groningen