Civil rights groups are raising concerns about the accuracy of the U.S. Census Bureau’s proposed new system for protecting privacy in a report released today. The report raises concerns about how the Bureau’s use of differential privacy as evidenced by recent demonstration data, will obfuscate the true population count of minority populations and frustrate compliance with federal (and some state) Voting Rights laws governing the redistricting process.
PRESS RELEASE April 5, 2021: Civil Rights Groups Issue New Report Detailing Potential Harm to Upcoming Redistricting Efforts
Highlight Concerns About the Census Bureau’s Differential Privacy Methodology
(Washington, D.C.) – Civil rights groups are raising concerns about the accuracy of the U.S. Census Bureau’s proposed new system for protecting privacy in a report released today.
The preliminary findings, released by MALDEF (Mexican American Legal Defense and Educational Fund) and Advancing Justice – AAJC (Asian Americans Advancing Justice – Asian American Justice Center), focus on a process known as “differential privacy” that introduces false information to help protect the confidentiality of information and privacy of people who responded to the 2020 Census.
The report warns that differential privacy may undermine the fitness of the data and affect some groups and areas more significantly, particularly minorities, including Latino and Asian Americans.
“Introducing false data to the Census cannot go forward if it leads to violations of federal protections of voting rights,” said Thomas A. Saenz, MALDEF president and general counsel. “Unfortunately, there are troubling indications in this report that the Census Bureau’s use of differential privacy on Census 2020 data releases could have just that result.”
These findings show that differential privacy, … is generally increasing the homogeneity of counties with clear majority race/ethnic populations. This could be alarming if this pattern of increasing homogeneity holds when looking at majority-minority districts drawn with these data products because this would indicate a likely increase in “false positive” majority-minority districts.
excerpt from the report
MALDEF and Advancing Justice – AAJC said the preliminary study, Impact of Differential Privacy & the 2020 Census on Latinos, Asian Americans, and Redistricting, reinforced concerns that data produced using differential privacy would not be of sufficient quality to meet the legal requisites of redistricting – the constitutional requirement of equal population and compliance with the federal Voting Rights Act of 1965 (VRA).
A decennial census is required by the U.S. Constitution. The data from the count is used to determine the number of seats each state will have in the U.S. House of Representatives. Census data are also used for purposes of redistricting, distribution of federal funding and assisting policymakers, businesses, and other interested stakeholders in assessing and addressing community needs.
The accuracy of the Bureau’s data analysis has implications for the ability of communities of color to achieve representation during redistricting, which is of crucial importance to MALDEF’s and Advancing Justice – AAJC’s work.
Both groups said the report’s findings indicate the Census Bureau must do more to meaningfully engage external stakeholders by providing more transparency about what has occurred with the development of differential privacy to date and the space and opportunity for stakeholders to inform the Bureau on how best to assess options for final decisions regarding differential privacy.
“Census data are vital to our communities, but our preliminary analysis of the Census Bureau’s demonstration products give us grave concerns about the negative effect the current differential privacy methods could have on redistricting. Unfortunately, the Census Bureau has primarily limited its engagement around differential privacy to an insular group of technical experts and failed to adequately engage data users earlier in this process,” said Terry Ao Minnis, co-author of the report and senior director of Census and Voting Programs at Advancing Justice – AAJC. “We produced this report to outline the concerns we have seen in the previous demonstration products. Now is the time for the Bureau to be transparent about the changes it is making and how it is addressing these concerns.”
MALDEF and Advancing Justice – AAJC said they will issue a final report analyzing the data to further understand the impact of differential privacy on redistricting later this month. MALDEF national redistricting coordinator Steven Ochoa is also a co-author of today’s report.
READ THE REPORT HERE
