Analysing biases in genealogies using demographic microsimulation

An incomplete understanding of biases affecting the representativeness of genealogies has hindered their full exploitation. We report on a series of experiments on synthetic populations designed to assess how different biases in ascendant genealogies can affect the accuracy of demographic estimates. Using the SOCSIM microsimulation programme and Swedish fertility and mortality data (1751-2022), we analyse three sources of bias: selection in direct lineages, incomplete reconstruction of family trees, and missing information on subpopulations. Comparing demographic measures derived from ‘fully-recorded’ and ‘bias-infused’ synthetic populations, we find that including only direct ancestors leads to underestimating total fertility rate (TFR) (c.a. −39%) before the fertility decline and overestimating life expectancy at birth (e0) (c.a. +42.2%) in the first two centuries. However, after including collateral kin, TFR underestimation was reduced to −2.4% and e0 overestimation limited to +1.5%. Our study shows that the completeness of family trees is essential for obtaining accurate demographic estimates.