Estimated reading time: 5 minutes

Today we released the second round of 2020 Census operational quality metrics. On April 26, we released the first round of quality metrics, along with the first counts from the 2020 Census. 

Operational quality metrics provide important information about the quality of the census by looking at how we obtained a response for each address.

In this blog, I provide highlights from today’s release much like we did from the first set released in our Examining Operational Quality Metrics blog. 

The first release of the operational quality metrics focused on how we resolved whether each address, including housing units and group quarters, was occupied, vacant, delete, or remained unresolved. This second release focuses on occupied and vacant housing units and the household size of occupied units. Occupied housing units are living quarters, such as a house, apartment or mobile home, with at least one person living or staying there.

In today’s release, we provide the metrics for single- and two-person households categorized by the census operation from which we obtained the response (such as from self-response, Nonresponse Followup, other census operations, or from count imputation), and the percentage of occupied and vacant housing units by census operation.

These metrics enable us to examine how data collection may have varied depending on the size of the household. As with our first release of operational quality metrics, it’s important to keep in mind that a difference among these metrics, across the two decades or among the states isn’t necessarily better or worse, but it’s useful information as we continue to examine census operations.

The metrics are available in an interactive dashboard and a downloadable table. The dashboard makes it easy to compare metrics across states and relative to national figures. (This dashboard is similar to, but separate from, the 2020 Census: Operational Quality Metrics dashboard released April 26.)

Comparable metrics from the 2010 Census are also available in both the dashboard and table.

The metrics from today’s release (as well as the April release) are available on the 2020 Census Data Quality webpage.

Here are a few takeaways:

• The average household size was similar between census operations and between the two decades. The average household size was 2.4 people for both households that self-responded to the 2020 Census and households enumerated in Nonresponse Followup. These averages were similar to the 2010 Census averages of 2.5 people per household in self-response and 2.6 people per household in Nonresponse Followup.
• Among households that self-responded, 26% of occupied households had one person and 35% had two people living there. Both are similar to corresponding portions from the 2010 Census (26% and 35%, respectively).
• Among households enumerated in Nonresponse Followup, 33% of occupied households had one person and 27% had two people living there. These values compare to 30% with one person and 27% with two people in the 2010 Census.
• Overall, 70% of all housing units in the 2020 Census (including occupied and vacant units) were counted by self-response. About 8% provided a self-response without a Census ID (“non-ID”), the unique number that links a response to the correct physical address. This was one of the innovations in the 2020 Census design, which provided flexibility so that people could respond to the census anytime, anywhere. In comparison, 63% of all occupied or vacant housing units in the 2010 Census were resolved by self-response and less than 0.5% of those without an ID, such as through a response on a “Be Counted” form.
• In the 2020 Census, 77% of occupied households were enumerated by self-response. This is higher than the 71% enumerated by self-response in the 2010 Census. Our research shows that self-responses tend to provide the best quality data on occupied households. Conversely, as expected, most vacant housing units were enumerated in Nonresponse Followup: 75% by household member or proxy respondent and 13% by administrative records in 2020, and 95% by household member or proxy in the 2010 Census (the use of administrative records was new for the 2020 Census).
  

From a demography standpoint, these preliminary numbers are the first information we’ve shared from the 2020 Census on household size. Additional data on household size will be available in future 2020 Census data products.

From a quality standpoint, the metrics show that the way we collected information for different-sized households is consistent with the 2010 Census, and the consistency suggests comparable quality. We’ll continue to examine these metrics in our operational assessments.

While no single number definitively quantifies census quality, examining these metrics and comparing them across geographies and with past censuses can shed light on how our operations and processes affect our ability to accurately count our nation’s population.

Providing these operational quality metrics is another example of the Census Bureau’s efforts to ensure high-quality 2020 Census counts. With unprecedented transparency, we are sharing the very metrics and methods that we use to review the 2020 Census, and we are planning to issue additional quality metrics alongside the release of the redistricting data.

Other ways that we are measuring the quality of the 2020 Census include:

• Comparing census results to 2020 Demographic Analysis estimates and other population benchmarks. (More information is available in our Comparisons to Benchmarks as a Measure of Quality blog.)
• Conducting a Post-Enumeration Survey to measure how many people and housing units were missed or counted erroneously in the census.
• Working with respected members of the scientific and statistical community to provide independent, external assessments of the census.
• Conducting a series of planned assessments and evaluations of the 2020 Census operations.

The Census Bureau is committed to sharing what we know, when we know it, to help the nation understand the quality of the 2020 Census results. We will continue to update the 2020 Census Data Quality webpage as new information becomes available.