ML01: Record Linkage Current and Future Advances to Inform Policy

Oregon Convention Center 

Linking data across various sources is crucial for producing accurate statistics and estimating effects of interventions. Declining survey response rates, technological advancements, and the abundance of data collected by different organizations underscore the importance of linking different datasets. Linkage algorithms attempt to merge information from public and private datasets to aid in obtaining more official accurate statistics, improve disease outbreak prediction, perform policy monitoring, and conduct comparative effectiveness research. However, the challenge in linking file commonly lies in the absence of direct identifying information across the different sources. Probabilistic and deterministic record linkage algorithms serve as powerful solutions that enable researchers to link entities across disparate sources even without unique identifiers like social security numbers or names. This roundtable aims to explore current and future research avenues in record linkage methods, covering topics such as, but not limited to, computational algorithms, software availability, privacy preservation, linkage error estimation and propagation, and the linkage of multiple data sources.

Record Linkage

Computational Methods

Error Propagation

Privacy Preserving 

Government Statistics Section

Record Linkage Interest Group