Record Linkage For Character-Based Surnames: Evidence from Chinese Exclusion

This paper proposes a novel pre-processing technique to improve record linkage for historical Chinese populations. Current matching approaches are relatively ineffective due to Chinese-specific naming conventions and enumeration errors. This paper develops a three-step process that both triples the match rate over baseline and improves match accuracy. The procedures developed in this paper can be applied in part or in full to other sources of historical data, and/or modified for use with other character-based languages such as Japanese. More broadly, this approach suggests the promise of language-specific linkage procedures to boost match rates for ethnic minority groups.