The Census Bureau is modernizing the way we will protect respondent privacy in 2020 Census data products. This first blog, part of a series planned to keep data users informed of progress and developments, focuses on ways that data users can get involved by providing feedback.
The Census Bureau’s mission is “to serve as the nation’s leading provider of quality data about its people and economy.” As an agency, we take this responsibility seriously. Census data are routinely used for critical decision-making at all levels of government, and they enable policymakers, businesses, analysts and researchers across the country to measure and assess trends about who we are and where we are going as a society.
As discussed in previous blogs, the Census Bureau has a legal obligation to protect the confidentiality of personal data, a major task in the face of 21st century privacy threats. To meet this challenge, we are modernizing our privacy protections for the 2020 Census by adopting differential privacy, a cutting-edge approach for ensuring confidentiality. As with all statistical techniques used to protect privacy, differential privacy will affect the quantity and quality of statistics we release. But we’re counting on input from our data users to help ensure our 2020 Census data products retain the same high quality and utility the nation has come to depend upon.
There’s no question Census Bureau employees are leaders in demography, geography, economics, statistics, computer science and many other fields. But that doesn’t mean we always know what’s “best” when it comes to the quality and usability of our data products. For that, we have always relied heavily on the advice, feedback and recommendations of our data users.
Last October in order to get users’ feedback, the Census Bureau released a set of 2010 demonstration data products generated by running 2010 Census data through an interim version of the 2020 Disclosure Avoidance System (DAS) that will be used for the 2020 Census. This allowed data users to compare the originally released 2010 Census data (protected using a legacy disclosure avoidance technique called “data swapping”) with the same data protected using differential privacy with specific, public parameter settings, and identify where and how the DAS may need to be improved to meet the public’s data quality needs.
In December 2019, the Committee on National Statistics of the National Academy of Sciences held a two-day workshop. During this workshop, data users from a wide variety of fields and disciplines presented findings from their evaluations of the 2010 demonstration data products. Some echoed conclusions of our own internal evaluations, others were new and have already begun informing development work to improve the design and optimization of the 2020 DAS.
We extend a sincere thank you to all of the data users who participated in the workshop or who have sent us feedback on the 2010 demonstration data products. As we continue to improve the 2020 DAS to ensure both the quality and confidentiality of 2020 Census data products, we will keep you posted with regular updates via this blog. As always, we welcome feedback and comments throughout this process.
The Committee on National Statistics is organizing follow-up expert meetings. We intend to use these meetings to document improvements and overcome remaining challenges in the 2020 DAS. We hope that data users contacted to participate will continue to be generous with their time and effort. We encourage you to send feedback to: