Dive Brief:
- The U.S. Census Bureau is modernizing its approach to data collection and privacy through a new Disclosure Avoidance System (DAS), which has raised questions about the 2020 census’ ultimate efficacy.
- The bureau is testing a "differential privacy approach," which uses both fake and real data to reflect the accuracy of population demographics while accounting for individual privacy. The approach uses "statistical noise to slightly alter data so that the link between the data and a specific person or business can’t be certain … the formulas allow us to balance between two opposing extremes: total accuracy and [total] privacy,” according to the bureau's website.
- The DAS "Draft Quality Measures and Benchmark" document is open for public comment until Friday, but a representative from the bureau told Smart Cities Dive it will consider suggestions submitted past that date. "We want to ensure we are measuring the attributes data users require to gauge the data's fitness for use," the representative said.
Dive Insight:
Andrew Reamer, a research professor at the George Washington Institute of Public Policy (GWIPP), told Smart Cities Dive whatever finalized privacy process the bureau chooses will have lasting implications on how census data can be used. While the 2020 DAS is solely optimized for the 2020 census, it will inform future developments of differential privacy to be fully implemented for data products like the American Community Survey (ACS) by 2025. Solutions will be "vetted by the scientific and user communities," to the likeness of stakeholders like population database IPUMS.
There’s concern that adding "noise," or synthetic data, in the way the DAS suggests could present inaccurate census data for rural areas and minorities. The bureau is working to answer questions throughout this modernization.
Data privacy is of utmost importance when collecting information for the census, as it is for collecting data in cities. Local governments are grappling with privacy and security concerns firsthand amid the novel coronavirus pandemic, with remote staff presenting IT challenges.
The pandemic has already presented challenges to the Census Bureau’s data collection work. Earlier this month, U.S. Secretary of Commerce Wilbur Ross and U.S. Census Bureau Director Steven Dillingham asked Congress for a 120-day extension for the delivery of final Census response counts. Field operations were suspended in the wake of COVID-19, and other deadlines for this year’s census have been pushed back.
While a delay might not greatly affect urban areas, the count of rural areas and groups at homeless shelters could be affected.
Albert Fontenot, associate director for decennial programs at the U.S. Census Bureau, said on a conference call last month that the pandemic shows how important the census truly is in ensuring communities receive funding for hospitals, schools and other infrastructure.
"Even though many things may seem uncertain at the moment, one thing isn't: the 2020 Census this year, it’s important to our nation that everyone respond," Fontenot said on the call. "The current situation underscores the need for census data."