Errors in 2020 census data due to new privacy controls: A user’s guide
Topics: Population Geography
, United States
, Spatial Analysis & Modeling
Keywords: 2020 census, error, uncertainty, population, neighborhoods
Session Type: Virtual Paper Abstract
Day: Monday
Session Start / End Time: 2/28/2022 11:20 AM (Eastern Time (US & Canada)) - 2/28/2022 12:40 PM (Eastern Time (US & Canada))
Room: Virtual 23
Authors:
Jonathan Schroeder, IPUMS, University of Minnesota
David Van Riper, IPUMS, University of Minnesota
Tracy Kugler, IPUMS, University of Minnesota
,
,
,
,
,
,
,
Abstract
For the 2020 U.S. Census, the Census Bureau adopted a new disclosure avoidance system that achieves “differential privacy” (a framework for quantifying the risk of disclosing an individual’s characteristics) by adding random errors into published data. The level of error varies among geographic levels and across tables, complicating efforts to assess the impact of errors for any particular analysis. To enable users to assess the distribution of errors, the Bureau has provided demonstration data generated by applying the 2020 disclosure avoidance system to 2010 data. Preliminary assessments of the demonstration data, along with the Bureau’s own summary metrics, indicate that substantial errors are not uncommon, especially for small areas and small population groups. While prior assessments have shed light on specific statistics and use cases, we aim here to present a general summary of data reliability across several commonly used levels, from states down to tracts and blocks, and across a wide range of characteristics, including race, Hispanic origin, group quarters type, and housing occupancy. We will provide a short list of guidelines and some simple, broadly applicable indices of reliability. Researchers, planners, policy analysts, and students will be able to use these guidelines and indices to assess the fitness of 2020 Census data for a range of uses.
Errors in 2020 census data due to new privacy controls: A user’s guide
Category
Virtual Paper Abstract
Description
This abstract is part of a session. Click here to view the session.
| Slides