Research

Preliminary Report: Impact of Differential Privacy & the 2020 Census on Latinos, Asian Americans and Redistricting

April 5, 2021

Preliminary findings reveal serious concerns about the impact of differential privacy as currently envisioned by the Bureau on our communities.

Group

Type:

Report

The Constitution requires a census every ten years in order to determine the number of seats each state will have in the U.S. House of Representatives. In addition to determining this reapportionment of House seats, Census data are also used for purposes of redistricting, distribution of federal funding, and assisting policy makers, businesses, and other interested stakeholders in assessing and addressing community needs. 

The Census Bureau has utilized different methods to meet its federal-law statutory mandate to protect respondents’ privacy and confidentiality in published statistics. For the 2020 census, the Census Bureau will use a new methodology for this purpose – “differential privacy.” Differential privacy is a mathematical method that uses statistical noise, or false information, to alter data so that the link between the data and specific persons or households cannot be ascertained. How the Census Bureau implements differential privacy could negatively affect civil rights enforcement for the next decade, including with respect to redistricting and voting rights.  

This preliminary report – the first such report from civil rights organizations on this topic – compares the demonstration products (which tested different configurations of differential privacy using 2010 Census data) to published 2010 Census data in an effort to assess the impact of differential privacy on a) total population and racial/ethnic populations and b) redistricting. 

This report seeks to answer some questions about how the different demonstration products met accuracy standards for total population among racial/ethnic groups and for purposes of redistricting, with an eye toward the fact that the window is closing for the Census Bureau to make specific decisions about how differential privacy will be implemented (by mid-2021). Our preliminary findings reveal serious concerns about the impact of differential privacy as currently envisioned by the Bureau on our communities'  ability to attain our fair share of political power, and surface the need to find solutions to improve the disclosure avoidance system for 2020 census data. 

Download: Preliminary Report District Data and Tables