On March 13, John Abowd, the Census Bureau’s Chief Scientist, and Victoria Velkoff, the Bureau’s Associate Director for Demographic Programs, published a blog post to give a status update on the Census Bureau’s strategy for avoiding disclosure of personal information as a result of data released from the 2020 Census. The Bureau had previously announced that it planned to implement the move to a standard called “differential privacy” (which uses an algorithm to injects precise amounts of random noise into data until it reaches a desired threshold of obfuscation). It released demonstration data products to give users a chance to see how 2010 Census data would have been affected under these conditions and sponsored a National Academies workshop to collect feedback (see COSSA’s article on this issue in ASA Footnotes).
In their blog post, Abowd and Velkoff note that feedback from the Academies workshop and elsewhere identified “unacceptable” errors created during the processing phase of the disclosure avoidance process: “To put it succinctly, the resounding message was that this interim version of the DAS [disclosure avoidance strategy] is generating significant error in the data that we need to resolve prior to the production of the 2020 Census Data Products.”
Abowd and Velkoff list several potential solutions that Census Bureau staff are exploring, including “changes to the geographic hierarchy used within the DAS, alternative estimation techniques to correct for the known biases of Non-Negative Least Squares optimization, and multiphase estimation of key statistics during post-processing.” They share that the Bureau is compiling “fitness-for-use” measures to evaluate the effectiveness of the solutions developed for different use-cases and situations. COSSA will continue to share information on these measures and solutions as details are released in the coming months.