Kinship study reveals universal patterns across human family trees
Analyzing 100+ datasets uncovers scale-free, small-world structures in genealogies
A team led by Japheth Carlson and Benjamin Webb at Brigham Young University has published a comprehensive analysis of genealogical networks, examining over 100 datasets from the Kinsources repository. The paper introduces "pseudogenerations," a novel method that infers temporal layers directly from network topology without requiring explicit birth or death dates. This allows researchers to study how family structures evolve over time across datasets that may have incomplete or inconsistent temporal metadata.
The results reveal several universal properties: genealogical networks exhibit scale-free-like degree and component-size distributions, meaning a few individuals have many connections while most have few. The networks also display small-world behavior, where most pairs of relatives are separated by surprisingly few marriage or genetic links. The authors identify 2-components (subgraphs where every pair of individuals is connected by at least two independent paths) as natural building blocks of family structure. They also document consistent disassortative mixing (people tend to marry those genetically dissimilar to themselves) and a strong bias toward unions at short genetic distances. Temporal shifts in recorded parental and child information suggest changing record-keeping practices over generations.
- Analyzed over 100 datasets from Kinsources using a new temporal inference method called 'pseudogenerations'
- Found scale-free degree distributions, small-world behavior, and disassortative mixing across all datasets
- 2-components emerge as fundamental structural units, and marriages show strong preference for short genetic distances
Why It Matters
Provides a universal framework for analyzing family history data, potentially improving ancestry predictions and database design for platforms like Ancestry or MyHeritage.