2020 Census data quality and accuracy
Due to the numerous changes to the census timeline that have led to questions about the quality of the enumeration process, a number of independent groups and advisory bodies have issued reports and recommendations on indicators that the Census Bureau could release in order to help establish confidence in the accuracy of the data. Below are some of those reports.
- Assessment of 2020 Census Data Quality Processes, JASON (independent scientific advisory group) report - February 23, 2021
- 2020 Census quality indicators: A report and updates from the American Statistical Association
- Reports from the Government Accountability Office
- December 9, 2020: The Bureau Concluded Field Work but Uncertainty about Data Quality and Accuracy, and Protection Remains
- December 3, 2020: Census Bureau Needs to Assess Data Quality Concerns Stemming from Recent Design Changes
- August 27, 2020: Recent Decision to Compress Census Timeframes Poses Additional Risks to an Accurate Count
- Testimony from House Oversight and Government Reform Hearing, December 3, 2020
- Evaluating the Accuracy of the Decennial Census: A Primer on the Fundamentals of Census Accuracy & Coverage Evaluation, Georgetown Center on Poverty & Inequality (November 2020)
- Other 2020 data for Washington state
- OFM's 2020 Estimates using 2010 Census as a base
- Census Bureau's Vintage 2020 Population Estimates for the United States and States
The Census Bureau has developed a new disclosure avoidance system referred to as Differential Privacy. Intended to protect the privacy of individual responses as required by Title 13. A number of data scientists have raised concern about the new system’s effect on data quality and usability. In a series of analyses, OFM has registered its concern about the bureau’s plan and is also closely following reports from other entities addressing data quality concerns.
- Briefing papers and perspectives
- The Challenge of Differential Privacy: Confidentiality vs. Usability - OFM, September 2020
- Implications of Differential Privacy for Reported Data on Children in the 2020 U.S. Census, William P. O'Hare, Consultant to Count All Kids 2020 Complete Count Committee
- Perspectives from the Policy Research Center of National Congress of American Indians
- Data, information and feedback
- April 28, 2021 Privacy-Protected Microdata Files
- OFM Feedback on April 28, 2021 Release of Privacy-Protected Microdata File - May 27, 2021
- Washington state excel comparison files for global privacy loss budget 12.2 (10.3 persons and 1.9 for housing units) [zip file: 22 MB in 11 files]
- Washington state CSV census block comparison files for privacy loss budget 12.2 (10.3 persons and 1.9 for housing units) [zip file: 20 MB in 12 files]
- Washington state excel comparison files for global privacy loss budget 4.5 (4.0 for persons and 0.5 for housing units) [zip file: 22 MB in 11 files]
- Privacy-Protected 2010 Census Demonstration Data | IPUMS NHGIS
- OFM Feedback on November 16, 2020 Release of Privacy-Protected Microdata File - Dec. 21, 2020
- Excel comparison files [zip file: 16 MB in 11 files]
- FSCPE Letter to Biden-Harris Transition Team - Nov. 24, 2020
- FSCPE Letter to the Data Stewardship Executive Policy Committee - Nov. 23, 2020
- OFM's 2020 Count Question Resolution Operation Federal Register Response - Oct. 2, 2020
- OFM Feedback on May 27, 2020 Release of Privacy-Protect Microdata File - Aug. 20, 2020
- Supporting charts [zip file]
- OFM Feedback on October 29, 2019 Release of Disclosure Avoidance System - Feb. 6 2020
- Excel comparison files [zip file: 84 MB in 15 files]
- Variables: differs by geodatabase size denoted below:
- Small: population, housing, occupancy and group quarters population [files size: 103 MB]
- Medium: same as "small" with the addition of population by race, population by total races tallied and population by sex by 5-year age group [file size: 131 MB]
- Large: same as "medium" with the addition of population by sex by 5-year age group by race and sex by 1-year age group (tract level only) [files size: 163 MB]
- Age sex charts
- Data sets for statistical analysis: SAS summary files (SF1 and DDP)
- CIC/FSCPE/SDC Feedback on Disclosure Avoidance System - Jan. 28, 2020
- Census Bureau: Data Protection and Privacy Program website
- Other places to find Census Bureau's 2020 Demonstration Files
- IPUMS and NHGIS website
Given the challenges from the 2020 census, a number of counties and cities have asked if there are mechanisms for requesting a correction to census data after results are released. The Census Bureau describes the generic procedure for requesting corrections across all of its data products per OMB guidelines.
In terms of the decennial census, the most frequent methods that cities and counties use are described below. The links refer to procedures and timelines as deployed in the 2010 census. Once information about these processes directly connected to the 2020 census is released, OFM will update the links.
- Count Question Resolution Program – State, local and Tribal area elected officials can challenge their jurisdiction’s 2020 Census counts through the Census Count Question Resolution (CQR) Program, which will accept challenge submissions from October 1, 2021 through June 30, 2023. The Bureau’s goal is to complete and adjudicate all challenges by September 30, 2023. The Census Bureau will not collect additional data during the challenge process, but the challenge could correct a geographic boundary, coding of a housing unit and housing units for a local area, or address processing errors. If a challenge results in a change, the Census Bureau issues official revised counts, which the Bureau and OFM will use in their subsequent estimate series.
- Request a Special Census – The Census Bureau may conduct a basic enumeration of population, housing units, and group quarters at the request of, and at the expense of, a governmental unit under the Special Census program. Local officials might request a Special Census when there has been a significant population change in their community due to growth or annexation. The Special Census program typically suspends operations two years before and two years after the decennial census.