Correlates of record linkage and estimating risks of non-linkage biases in business datasets

Abstract

Researchers often utilise datasets that link information from multiple sources, but non-linkage biases caused by linked and non-linked subject differences are little understood, especially in business datasets. We address these knowledge gaps by studying biases in linkable 2010 UK Small Business Survey datasets. We identify correlates of business linkage propensity, and also for the first time its components: consent to linkage and register identifier appendability. As well, we take a novel approach to evaluating non-linkage bias risks, by computing dataset representativeness indicators (comparable, decomposable sample-subset similarity measures). We find that the main impacts on linkage propensities and bias risks are due to consenter / non-consenter differences explicable given business survey response processes, and differences between subjects with and without identifiers caused by register under-coverage of very small businesses. We then discuss consequences for the analysis of linked business datasets, and implications of the evaluation methods we introduce for linked dataset producers and users.

More information

Identifiers

ORCID for Gabriele Durrant: orcid.org/0009-0001-3436-1512

ORCID for Peter W F Smith: orcid.org/0000-0003-4423-5410

Catalogue record

Export record

Altmetrics