When we collected data for the 2020 Census, we assured people that their responses would be kept confidential as required by law.

Because the redistricting data have demographic characteristics available for very small areas, it’s essential we take steps to protect data reported by or on behalf of individuals in our published statistics.

The redistricting data will be the first 2020 Census data protected using differential privacy. This modern method protects individual information while letting us share important statistics about communities. Differential privacy has been successfully used in several other Census Bureau data products. We recently shared more about the need for differential privacy in the Modernizing Privacy Protections for the 2020 Census: Next Steps blog and in the Protecting Privacy in Census Bureau Statistics video.

By design, we’ve carefully calibrated how much protection or noise to add so that the results strike a balance between data protection and precision. More information is available in our recent Key Parameters Set to Protect Privacy in 2020 Census Results statement.

With these parameters, some small areas like census blocks may look “fuzzy,” meaning that the data for a particular block may not seem correct. Importantly, our approach yields high quality data as users combine these "fuzzy” blocks to form more significant geographic units like census tracts, cities, voting districts, counties, and American Indian/Alaska Native tribal areas. Our calibration was designed to achieve acceptable quality thresholds for these levels of geography.

So, if you’re looking at block-level data, you may notice situations like the following:

• Occupancy status doesn’t match population counts. Some blocks may show that the housing units are all occupied, but the population count is zero. Other blocks may show the reverse: the housing units are vacant, but the population count is greater than zero.
• Children appear to live alone. Some blocks may show a population count for people under age 18 but show no people age 18 and older.
• Households appear unusually large. For example, you may find blocks with 45 people, but only three housing units.

Though unusual, situations like these in the data help confirm that confidentiality is being protected.

Noise in the block-level data will require a shift in how some data users typically approach using these census data.

Instead of looking for precision in an individual block, we strongly encourage data users to aggregate, or group, blocks together. As blocks are grouped together, the fuzziness disappears. And when you step back with more blocks in view, the details add together and make a sharp picture. 

Finally, I’d like to say a little more about data quality — something in which we know there is keen interest given the challenges presented in 2020. 

As always, the Census Bureau is committed to being transparent about the quality of the 2020 Census results. Shortly after the August release of redistricting data, we plan to release additional operational quality metrics to give further insight into how we collected 2020 Census responses and what that might mean for the quality of the data.

Also, data users will be able to compare for themselves how the local and demographic 2020 Census counts compare against other population benchmarks, such as our annual population estimates.

Throughout data processing, we’ve been comparing the counts to benchmarks as part of our quality checks on the data. We have conducted one of the most comprehensive reviews in recent census history.

As part of our review, we’ve had to contend with higher item nonresponse rates for characteristics than we’ve experienced in past censuses. This means while people were counted, some people left one or more questions blank, even if they completed most of the census questionnaire.

We saw this for almost every variable in the questionnaire — age, sex, race, Hispanic origin, and whether the home is owned or rented. We saw this across the board regardless of whether the response was submitted online, by paper, over the phone or during in-person interviews with our enumerators. These blank responses left holes in the data which we had to fill. We’re still analyzing this and will share how we addressed this challenge in an upcoming blog.

In addition, please remember the 2020 Census represents a count of everyone living in the nation on April 1, 2020. This is important to note when comparing to other Census Bureau surveys or other non-Census Bureau data sources.

Where there are questions about the data, we have a process — the Count Question Resolution program. This program is limited in scope and won’t change either the redistricting data or the apportionment results, but it will help us update the census totals we use to build our population estimates each year. We’ll share more details about this program later.

Every way we’ve analyzed the 2020 Census — through our extensive reviews during data processing, by comparing the numbers to population benchmarks, and looking at the operations — the census data are high quality and are fit to use for redistricting. In fact, the quality of the 2020 Census data is quite remarkable amid all the challenges we faced last year.