Record Linkage

NORC has conducted dozens of record linkage projects for a wide variety of clients, including linking National Hospital Care Survey Data to the National Death Index and CMS Data for the National Center for Health Statistics and linking Survey of Occupational Injuries and Illnesses to Occupational Safety and Health Administration ITA data for the Bureau of Labor Statistics.

Methods

NORC's approach mainly based on the Fellegi-Sunter record linkage paradigm, with some linkages developed as machine learning applications with a high degree of customization and enhancement. The enhanced methods we use include:

  • Methods to estimate linkage error
  • Developing synthetic data to represent linked data elements
  • Privacy preserving record linkage (PPRL)
  • Linkage routine optimization: for accuracy and speed
  • Use of custom code, propriety code modules, and commercial off-the-shelf (COTS) software

Publications

  1. Resnick, Dean, Lisa Mirel, Marc Roemer, and Scott Campbell. "Adjusting Match Weights to Partial Levels of String Agreement in Data Linkage." ASA Proceedings (2020) of the Joint Statistical Meetings, American Statistical Association (Alexandria, VA).
  2. Asher, Jana, Resnick, Dean, et al. "An Introduction to Probabilistic Record Linkage with a Focus on Linkage Processing for WTC Registries." International Journal of Environmental Research and Public Health 17.18 (2020): 6937.
  3. Datta, A. Rupa, Gabriel Ugarte, and Dean Resnick. "Linking Survey Data with Commercial or Administrative Data for Data Quality Assessment." Big Data Meets Survey Science: A Collection of Innovative Methods (2020): 99-129.
  4. Resnick, Dean, and Jana Asher. "Measurement of Type I and Type II Record Linkage Error." ASA Proceedings (2019) of the Joint Statistical Meetings, American Statistical Association (Alexandria, VA).
  5. Dean Resnick. Book Review: "Register-based Statistics". Journal of Official Statistics. Volume 31, Issue 1, Pages 141–142, ISSN (Online) 2001-7367, DOI: 10.1515/jos-2015-0008, March 2015.
  6. Resnick, Dean, "The Estimation of Match Validity under the Fellegi-Sunter Paradigm without Assuming Identifier-Agreement Independence." ASA Proceedings (2017) of the Joint Statistical Meetings, American Statistical Association (Alexandria, VA).
  7. Resnick, Dean M., Christine S. Cox, and Lisa B. Mirel. "Using synthetic data to replace linkage derived elements: a case study." Health Services and Outcomes Research Methodology (2021): 1-18.
  8. Campbell, Scott R., Dean M. Resnick, Christine S. Cox, Lisa B. Mirel. "Using Supervised Machine Learning to Identify Efficient Blocking Schemes for Record Linkage." Accepted for publication in Statistical Journal of the IAOS (International Association for Official Statistics).