By Wendy Underhill
Privacy is the hot new topic: It’s one of NCSL’s priorities, The New York Times has a deep privacy project underway, and the ubiquity of smart devices has made data privacy a personal and state concern.
But for the U.S. Census Bureau, privacy is a hot old topic. The bureau has been required to keep individual respondents’ information private all along. For decades, the bureau has either not released data from small geographic units that might be used to identify individuals, particularly those from small racial and ethnic minorities. Since 2000, the bureau has used “data swapping” between census blocks as its main disclosure avoidance technique. (The census block is the smallest unit of geography maintained by the bureau.)
For 2020, the bureau will use something called “differential privacy,” a method that “injects noise” (that’s a phrase straight from the bureau) at all geographic levels with the exception of state total population, in order to make it harder to cross-reference census data products with each other, or with data available elsewhere.
And yet, the census requirements for an accurate count and the protection of respondents and their data create a natural tension: The more accurate (and therefore usable) the reported data is, the easier it may be to identify individual responses. But as the raw data is altered before being reported (to protect confidentiality), the less usable the publicly released data is.
Who cares? Anyone who uses census data: redistricters, policymakers, budgeteers.
Read all about it (or at least a bit about it) on NCSL’s new webpage, Differential Privacy for Census Data Explained.
Wendy Underhill is the director of NCSL’s Elections and Redistricting Program.