A Simulated Reconstruction and Reidentification Attack on the 2010 U.S. Census
Abstract
We show that individual, confidential microdata records from the 2010 U.S. Census of Population and Housing can be accurately reconstructed from the published tabular summaries. Ninety-seven million person records (every resident in 70% of all census blocks) are exactly reconstructed with provable certainty using only public information. We further show that a hypothetical attacker using our methods can reidentify with 95% accuracy population unique individuals who are perfectly reconstructed and not in the modal race and ethnicity category in their census block (3.4 million persons)--a result that is only possible because their confidential records were used in the published tabulations. Finally, we show that the methods used for the 2020 Census, based on a differential privacy framework, provide better protection against this type of attack, with better published data accuracy, than feasible alternatives.
Cite
@article{arxiv.2312.11283,
title = {A Simulated Reconstruction and Reidentification Attack on the 2010 U.S. Census},
author = {John M. Abowd and Tamara Adams and Robert Ashmead and David Darais and Sourya Dey and Simson L. Garfinkel and Nathan Goldschlag and Michael B. Hawes and Daniel Kifer and Philip Leclerc and Ethan Lew and Scott Moore and Rolando A. Rodríguez and Ramy N. Tadros and Lars Vilhuber},
journal= {arXiv preprint arXiv:2312.11283},
year = {2025}
}
Comments
This is the accepted Harvard Data Science Review (2025) paper. The accepted supplemental text is here: https://arxiv.org/abs/2312.11283v2