A First Full-Count Linking of English and Welsh Censuses, 1851-1921

Abstract

Historical census linking can provide researchers with new evidence for much finer studies of population and economy, turning cross-sectional observations into longitudinal data by tracking individuals and their changing characteristics across censuses. In this paper, we adapt an unsupervised Bayesian census linking method to British censuses between 1851 and 1921. We split the process into several phases, which allows us to reduce the amount of blocking required and minimize false negatives. We also combine immutable and mutable characteristics to disambiguate cases with a high level of similarity. Finally, by incorporating the entire set of marriage civil registration, we are able to halve the gender bias compared to existing methods in the field. We provide researchers with the first full-count linking for England and Wales 1851-1921 together with detailed guidelines on how to make the best use of them for future research based on a detailed description of possible biases in the linked data.