Confidentiality and Privacy Protection after Record Linkage: Laying the Groundwork for Synthetic Record Linkage

Yajuan Si
Research Associate Professor, Survey Research Center, Institute for Social Research

Co-Investigators

  • David S. Johnson

Abstract

Augmenting surveys with information about the places sample members have lived can enhance the value of survey data. Restricted enclaves with remote access are valuable for sharing such data, yet barriers to use outside such environments remain because of confidentiality and privacy concerns. The Census Bureau has recently taken steps to create synthetic versions of some of its data products, but the robustness of such data to answer questions not explicitly considered by the synthetic data generation technique has been questioned. Using the Panel Study of Income Dynamics, this pilot project will develop and evaluate procedures to create survey data linked with synthetic geographic data designed to address confidential and privacy as well as analytic concerns.

Outcomes

  • National Science Foundation. 2217456. Enhancing Synthetic Data Techniques for Practical Applications. Role: Co-PI.
  • Mathur, S, Si, Y and Reiter, JP (2024). "Fully synthetic data for complex surveys," Survey Methodology, forthcoming. Preprint available: https://arxiv.org/pdf/2309.09115
  • Si, Y. Fully Synthetic Data for Complex Surveys. Federal Committee on Statistical Methodology (FCSM) Research & Policy Conference, College Park, MD, Oct. 2023
  • Si, Y and Johnson, D. Confidentiality and Privacy Protection after Record Linkage: Laying the Groundwork for Synthetic Record Linkage. 4th Annual MiCDA Pilot Project Virtual Symposium. May 18, 2021