This project is undertaking research that will improve Census capabilities for entity resolution and record linkage. We have brought together world-class researchers from multiple disciplinary backgrounds, including statistics, survey methodology, economics, history, computer science, and demography. Through this work, we will:
- build on ongoing projects linking Census data products to themselves, to external surveys, and to administrative data
- develop and evaluate approaches that show substantial improvements over current best practices in record linkage and entity resolution
- produce usable and useful diagnostics that can provide substantial improvements in record linkage assessment, with metrics of quality that capture bias and uncertainty
As part of a collaboration with the U.S. Census Bureau, the University of Michigan Institute for Social Research (ISR) has started a working group on record linkage. Join us as ISR researchers and others discuss their ongoing projects on record linkage and related topics.
Past Seminars
January 24, 2022
- Data Analysis After Record Linkage: Sources of Error, Consequences, and Possible Solutions
- Brady West (University of Michigan) and Martin Slawski (George Mason University)
Slides
February 21, 2022
- Scaling Record Linkage
- H.V. Jagadish (University of Michigan)
Recording, Slides
March 21, 2022
- Modernizing Person-Level Entity Resolution with Biometrically Linked Records
- Michael Mueller-Smith (University of Michigan) and Matthew Gross (Twitter)
Recording, Slides
April 18, 2022
- Finding Needles in Haystacks: Multiple-Imputation Record Linkage Using Machine Learning
- Dhiren Patki et al (Federal Reserve Bank of Boston, Census Bureau, & University of Michigan)
Recording, Slides
May 16, 2022
- Decennial Census Digitization and Linkage Project
- Trent Alexander (University of Michigan) and Katie Genadek (Census Bureau)
Recording, Slides
