What you need to know about the 2020 American Community Survey

In October, the U.S. Census Bureau released a report detailing why it planned to release the results of the 2020 American Community Survey (ACS) 1-year estimates in an experimental format. The ACS is the only annually updated source of detailed social, demographic, and economic information for all communities in the United States.

In addition, the Bureau announced in November that it would delay the release of the 5-Year ACS estimates that had been planned for December 2021. At present, the Bureau plans to release data for the 5-Year ACS in March 2022 but has not confirmed that it will be a standard release. They plan to provide more information in December.

This blog post details the planned changes for both the 1-Year ACS and the 5-Year ACS and what we know so far. We will continue to update you as we learn more about these changes.

Why the American Community Survey is important

The American Community Survey (ACS) is an ongoing survey that is sent to a sample of households across the United States and Puerto Rico on a monthly basis (~3.5M households each year). The survey asks about topics that aren’t on the census, such as: educational attainment, Internet access, transportation, and employment. The ACS data is released annually for all geographies:

• 1-year estimates are available for areas with populations of 65,000 or more;
• 5-year estimates are available for smaller areas (these combine data from the most recent five years to create more reliable estimates for small geographies)

The ACS is best used for understanding population characteristics—percents, means, medians, and rates—rather than population counts.

Every year, the Bureau releases both summary data—pre-tabulated tables and profiles—and Public Use Microdata Sample (PUMS) with individual-level responses for the 1-Year and 5-Year ACS.

Why the 5-Year ACS is important

The 5-Year ACS estimates provide detailed information for all geographies (down to the block group level). These combine the last five years of American Community Survey data for more accurate estimates of small areas. For example, the 2019 5-Year ACS uses data collected from 2015 through 2019.

Decisions made about how to incorporate the 2020 data into the 5-Year file will impact data users for the next five years, as the 2020 data will be a part of each 5-Year file until 2026 when the 2021-2025 5-Year file will be released.

The effect of COVID-19 on data collection in 2020

The COVID-19 pandemic significantly disrupted data collection for the ACS in the following ways:

• The number of interviews completed in 2020 was 32% smaller than in 2019
• The sample had significant nonresponse bias, meaning the characteristics of individuals who responded to the survey were not broadly representative of the population: respondents were more likely to be White, have a college education, live in a single-family house, and earn higher incomes.

After evaluating the data collected, the Bureau determined that the 2020 ACS data did not meet their Statistical Quality Standards. Shortly after, they delayed the release of the ACS data products and modified the release plan for the 1-Year ACS.

Changes to the 1-Year American Community Survey

Due to the impact of COVID-19, the Bureau is not releasing its standard set of 1-Year ACS estimates and made the following changes:

• Data is experimental. The nonresponse bias in the data collected means that the Bureau could not use its typical approach to weighting the data to create population totals. Instead, they developed a new method to weight the data and the results may still not meet all the Bureau’s data quality standards. As such, it is labeled an “experimental” data product.
• Data release is delayed. The data release was postponed from mid-September to late November. The experimental summary data will be released by November 30, 2021. The PUMS microdata with experimental weights will be released in December.
• Limited set of data available. Typically, the Bureau releases more than 1,350 tables in the summary file, in addition to other tables and profiles; the experimental data release will contain a set of 54 tables.
• Limited set of geographies. Summary data tables will only be available for the nation, states, and the District of Columbia, instead of all geographies with a population of 65,000 or more.
• Data will only be available as an Excel download. Data will not be available through the Bureau’s common dissemination mechanisms: data.census.gov and the Application Programming Interface (API).

How to work with the 2020 1-Year Experimental ACS data

For individuals who use ACS data, the Bureau makes the following recommendations for working with the 2020 Experimental ACS data:

• Do not make comparisons to prior ACS years, i.e., don’t compare 2020 to 2019
• Use the experimental data with caution
• Consider using 2019 1-Year ACS or 5-Year ACS data or look to alternate sources if you need to use the data to allocate funds, evaluate programs, or make decisions

What this means for future ACS releases

The data being collected in 2021 is less impacted by the pandemic and the Bureau anticipates that they will be able to resume regular release schedules. However, they will not be able to make this determination until all data are received and processed (late spring 2022).

The 2021 ACS data will also be the first ACS data to use the new 2020 Census results and 2020 Census geographies, so ACS data users should expect population totals and geographic boundaries to change.

We’ll be closely monitoring what the Bureau says over the next several weeks and will provide updates as we learn more.  If you have specific questions, please let us know at demography@unc.edu and we’ll do our best to answer them.

Republish our content for free under a Creative Commons license.

Need help understanding population change and its impacts on your community or business? Carolina Demography offers demographic research tailored to your needs.

Contact us today for a free initial consultation.

Categories: Carolina Demographics, Census 2020