June 10, 2024: The U.S. Census Bureau has released new guidance to help data users assess whether the 2020 Census data is fit-for-use for their specific individual applications and use cases considering the measures taken for disclosure avoidance in the 2020 Census. 

This new guidance and accompanying resources include:

• measures of the comparative impact of different degrees of aggregation to reduce disclosure avoidance-related error,

• best practices for calculating averages and ratios for sub-state geographies, and

• an innovative methodological approach for estimating confidence intervals (margins of differential privacy-related error) for data protected using the TopDown Algorithm. At present, the validity of this approach is empirically demonstrated using 2010 Census data. Resources to estimate these confidence intervals for the 2020 Census Redistricting (P.L. 94-171) Summary File, 2020 Census Demographic and Housing Characteristics File (DHC), and any custom tabulation generated from the forthcoming 2020 Census Privacy-Protected Microdata File (PPMF), will be released later this year in conjunction with the public release of the 2020 PPMF. 

The release is part of the Census Bureau’s ongoing commitment to help data users effectively use differential privacy-protected data in statistical and demographic analysis.

The Census Bureau plans to release tutorials to help interested data users understand how to incorporate this guidance and calculate the confidence intervals. We’ll provide more information at a later date. In the interim, please address any questions to us at 2020DAS@census.gov. 

Today's Links:

• Fact Sheet: Calculating and Interpreting Average Household Size Ratios in the Demographic and Housing Characteristics File 

• Fact Sheet: Assessing Disclosure Avoidance Uncertainty in the 2020 Census: Determining Reliability Thresholds for Demographic and Housing Characteristics Data

• Estimating Confidence Intervals for 2020 Census Statistics Using and Approximate Monte Carlo Simulation (hosted on the Registry for Open Data on AWS)