Arbeitsbereich

Bevölkerungsdynamik und Nachhaltiges Wohlbefinden

Auf einen Blick Projekte Publikationen Team

Projekt

Predicting Work-Family Life Course Sequences

Linda Vecgaile, Emilio Zagheni, Luca Badolato (The Ohio State University, Columbus, Vereinigte Staaten); in Zusammenarbeit mit Luiz Felipe Vecchietti (Institute for Basic Science, Daejeon, Korea, Süd)

Ausführliche Beschreibung

As societies age and governments raise statutory retirement ages, maintaining employment in later life becomes both more necessary and more challenging. Disruptions such as job losses, caregiving demands, or health declines can push older workers out of the labor market at precisely the time when they are expected to work longer. It is therefore increasingly important to understand why some people experience stable late-career pathways while others move into trajectories of insecurity, and how unfavorable life course developments might be prevented.

Against this background, this project combines life course theory with state-of-the-art machine learning methods to both predict and explain life course trajectories. By bringing advances from contemporary machine learning approaches into sociology and public policy research, we show how tools like sequence modeling, which was originally developed in areas such as natural language processing , can be adapted for studying ordered life course states and transitions. In addition, we explore counterfactual “what if” analyses that ask how predicted trajectories shift under alternative life course scenarios and potential interventions. The objective is to identify social groups across institutional contexts who are at risk of employment or health hardships, and to generate insights into the mechanisms that drive divergent pathways, in order to support evidence-informed prevention and intervention design.

As an initial application within the project, one study applies a probabilistic sequence modeling framework to predict late-career employment trajectories, focusing not only on the most likely future pathway, but also on credible alternative pathways derived from the model’s predicted probability distributions. This moves life course research beyond single “best guess” forecasts by making uncertainty explicit and showing where multiple futures remain plausible – an approach that remains uncommon in sociological sequence analysis. We identify seven characteristic late-career trajectory types ranging from stable employment to more precarious pathways (see figure below). A key finding is that, for many individuals, several future paths remain plausible, and the degree to which these probabilities are concentrated or diffuse is itself substantively meaningful: Some trajectories are highly stable (low uncertainty), while others show substantial branching (high uncertainty), thus highlighting where timely support could help steer pathways toward more favorable outcomes.

Another strand of this project connects late-career employment trajectories to women’s earlier work-family histories, and asks what “longer working lives” may imply for health. As many countries encourage higher female labor force participation to address labor shortages, a key concern is whether more time spent in employment comes with cumulative strain, especially when paid work is added to persistent caregiving responsibilities. Using life course trajectory approaches combined with causal “what if” analyses, this research shows that the health consequences of following typical work-family pathways are shaped by the institutional context: Supportive welfare arrangements can buffer risks, while less supportive contexts may amplify penalties and widen health inequalities in later life.

Publikationen

Vecgaile, L.:

The timing of poverty prevention: how early can women’s near-retirement poverty risk be detected?

MPIDR Working Paper WP-2026-031. (2026)    

Vecgaile, L.:

Convergence and divergence: from life-course trajectories to age 50 to projections of gendered old-age income inequality in Germany

MPIDR Working Paper WP-2026-032. (2026)    

Vecgaile, L.; Spata, A.; Vecchietti, L. F.; Zagheni, E.:

Predicting individual life trajectories: addressing uncertainty in social employment transitions

SocArXiv papers. unpublished. (2025)