Frequently Asked Questions

Who should use this tool?

This tool was developed for planners, administrators, policymakers, teachers and students, reporters, researchers, and anyone interested in census data.

Why should I use this tool?

This tool exists to assist a census data user in identifying deviations between expected population counts and the 2020 Census Redistricting Data file, and offers contextual resources for each county showing the atypical outcomes from 2020. The tool compiles this information into a downloadable report and points users to additional local data sources or experts to seek more assistance.

What motivated this tool?

In a traditional decennial census year, known errors in data can come from a variety of sources, including enumeration difficulties in hard-to-count communities, issues with the collection instrument (e.g., mail-in or electronic form), faulty or incomplete address lists, and methods used to protect respondent privacy. These sources of error are fairly well-documented, and the impacts can be quantified within a few degrees of accuracy. However, the 2020 Census was subject to these potential errors as well as a number of additional sources of known and unknown errors:

• COVID-19 impacts:
  • Local stay-at-home orders
  • The displacement of college students
  • Changes to pre-trial release at jails
  • Prison sentence commutation
  • Disproportionate deaths in nursing homes
  • Federal, state and local eviction moratoriums
  • A delayed in-person enumeration process
• Survey instrument (internet self response)
• Natural disasters such as hurricanes, wildfires, and more.
• A new statistical method for protecting privacy which adds noise to the data
• Government interference
  • Executive orders to collect information about legal status in the U.S.
  • Executive orders affecting collection deadlines
  • A decrease in government trust across multiple demographic and social groups

This tool provides census counts alongside contextual information that shows trends during 2020 Census data collection, including how response rates, local economic conditions, and COVID-19 case counts fluctuated during 2020.

What can this tool reveal, and what can it not reveal?

This tool aggregates contextual information about a selected county to identify potential sources of error or bias in the 2020 Census. It does not reveal where or what the potential undercounts or overcounts are, how federal funding may have been impacted by these factors, or attempt to measure the effects of errors. This tool presents contextual information with a measure of divergence between expected population counts versus actual counts.

What is a measure of divergence?

Each year, the Census Bureau produces a set of population estimates based on the previous decennial census and various components of change (e.g., births, deaths, and migration). The Census Bureau produces population, housing, and group quarters estimates. These can be compared to the actual counts released by the Bureau in a decennial census year.

How are measures of divergence calculated?

We measured divergence using the 2010 census as the “gold standard.” First, we pulled in the PL 94-171 and estimated data from 2010 and 2020, and proportional differences between counts and estimates were calculated for each measure. We then classified each county based on its population size. We labelled counties below 40,000 population as “small”, counties between 40,000 and 100,000 as “medium”, and counties above 100,000 in population as “large.” 

We calculated the mean and standard deviations by the county size and measure category to create the 2010 cutpoints. We next compared the 2020 proportions to these cutpoints to establish whether the 2020 divergence was close to expectation (within 1 standard deviation), slightly divergent from expectation (between 1 and 2 standard deviations), or highly divergent from expectation (over 2 standard deviations).

What do I do if my county is divergent from expectation?

Changes between population estimates and actual counts — which we refer to as the divergence from expectation — can be attributed to errors in the previous decennial census (2010 Census), errors in the most recent census (2020 Census), or errors in the components of change (more births, deaths, or migration than anticipated). If the selected county is highly divergent from expectation, we recommend reaching out to your local data experts, usually a State Data Center or Federal-State Cooperative for Population Estimates (FSCPE) contact, to learn more about the reliability of the counts and components of change, or file an appeal to the Census Bureau under their (yet to be determined) new rules about count resolution.

What data was used to create this tool?

All of the data used to develop the divergent measures was obtained from the U.S. Census Bureau. Contextual data came from the U.S. Census Bureau, the Bureau of Labor Statistics, FEMA, and the New York Times. All of the data sources and methods are outlined in our documentation.

How do I download this?

The data and workbook are available for download through our GitHub repository.

What are the next steps for the dashboard?

As the Census Bureau plans on releasing additional data, including the Demographic Profiles, the Demographic and Housing Characteristics file, and the Detailed Demographic and Housing Characteristics file, we will continue updating the 2020 Census County Assessment Tool to reflect new information. We also plan to add analytic abilities and allow users to drill down into Metropolitan Statistical Areas and census tracts.

Who created this tool?

This tool was created by the Massive Data Institute at Georgetown University’s McCourt School of Public Policy in collaboration with partners at the Urban Institute, with generous funding from the Tableau Foundation and the Alfred P. Sloan Foundation.