Presented by the Committee on Applied Demography
The U.S. Census Bureau implemented a new Disclosure Avoidance System (DAS) based on differential privacy (DP) to protect respondent information in the published 2020 Census data products. The 2020 Census DAS protects confidentiality by infusing small amounts of "noise" (error) into published tabulations. While the magnitude of this privacy-preserving error is generally small, it can impact fitness-for-use of the resulting data for particular use cases, especially for analyses focusing on very small geographic areas or demographic subgroups.
This session will provide census data users with the information and resources they need to be able to assess the fitness-for-use of 2020 Census data for their specific, individual applications and use cases, and to be able to adapt their statistical analyses accordingly. Special attention will be paid to the comparative magnitude of DP noise vs. other known sources of error (e.g., coverage and operational error), the known properties of DP noise distributions, and the sources and implications of biases that can arise from the 2020 Census DAS post-processing of the privacy-protected data prior to tabulation. The guidance and resources on assessing fitness-for-use provided in this course will cover major demographic and statistical use cases for both person and household data from the 2020 Census Demographic and Housing Characteristics (DHC) file (the 2020 Census successor file to the SF1 from prior censuses), and for both published and user-defined geographic areas.
Read full proposal.
Add to registration.