January 29, 2024: Today, the U.S. Census Bureau released the Detailed DHC-B Proof of Concept for the 2020 Census Detailed Demographic and Housing Characteristics File B (Detailed DHC-B) for public review and feedback.

The Detailed DHC-B includes household type and tenure data for approximately 1,500 detailed racial and ethnic groups and American Indian and Alaska Native tribes and villages. This proof of concept illustrates the proposed 2020 Census data product design and disclosure avoidance settings using 2010 Census data. 

The Census Bureau is also hosting a webinar today at 3:00 p.m. ET to guide you through the proof of concept and address any questions you might have. Refer to the webinar information below for details. 

A 30-day public feedback period on the proof of concept concludes on Wednesday, February 28, 2024. 

• Topics: Household type (e.g., family and nonfamily) and tenure (i.e., owner or renter occupied) data, including total household counts by householder race, ethnicity, or American Indian or Alaska Native (AIAN) tribe or village.
• Groups: Improvements in questionnaire design, processing, and coding allow for the release of approximately 1,500 detailed racial and ethnic groups and AIAN tribes and villages. This includes tables for both detailed groups (disaggregated groups such as Puerto Rican, the Navajo Nation, Samoan, etc.), and regional groups (such as Caribbean Hispanic, American Indian, Polynesian, etc.).
• Geographies: Nation, states, counties, census tracts, places, and American Indian/Alaska Native/Native Hawaiian (AIANNH) areas.  
• Companion file to the Detailed DHC-A: The Detailed DHC-B is the companion file to the Detailed DHC-A. The Detailed DHC-A provides population counts and age and sex statistics for the same detailed groups and geographies; it was released in September 2023.
• Data availability and adaptive design: The design ensures that only statistics that meet our accuracy and confidentiality standards are released. Adaptive design provides more granular statistics for larger populations and less granular statistics for smaller populations based on population thresholds for a given geography.
• Planned release: September 2024.

The Detailed DHC-B Proof of Concept includes descriptions of what data users can expect to encounter in the 2020 Detailed DHC-B, including:

• Detailed DHC-B Proof of Concept.
  • Table content and geographic availability.
  • Amount of data available for detailed and regional groups, how data are processed, accuracy of data, and data limitations.
• Detailed Summary Metrics.
• Tenure Examples resource.
• Graphics: Adaptive Design Process and Adaptive Design Summary.
• Privacy-Loss Budget Allocation resource.
• Frequently Asked Questions (FAQ) resource.
• The proof of concept does not include accompanying summary files or a full run of the 2010 Census data. Because this product is based on 2010 Census data, the increase in the amount of detailed race and ethnicity data that will be available in the 2020 Census Detailed DHC-B is not reflected in the proof of concept.

The amount of data a group receives is dependent on the Detailed DHC-A total population count. Groups that are eligible to receive Detailed DHC-B household type and tenure data receive one of four household type tables and one of two tenure tables. The tables vary by the amount of data, but they build upon each other. All tables include total household count. Refer to the Adaptive Design Summary graphic for more information.

The Detailed DHC-B is the companion product to the Detailed DHC-A. Both products provide data for detailed racial and ethnic groups and American Indian and Alaska Native (AIAN) tribes and villages. The Detailed DHC-A provides total population and sex by age data, while the Detailed DHC-B provides household type and tenure data. The Detailed DHC-A and Detailed DHC-B use algorithms (SafeTab-P and SafeTab-H, respectively) that produce privacy-protected tabulations directly from the confidential data. The Detailed DHC-A was released in September 2023, and the Detailed DHC-B is set for release in September 2024. Additional information is in the Detailed DHC-B Proof of Concept.

Common to Both the Detailed DHC-A and Detailed DHC-B:

• Disclosure avoidance noise is “tuned” to meet preset margins of error (MOEs). This allows data users to more easily analyze accuracy rather than having to research the MOEs across population groups, tables, and geographies. MOEs reflect the noise-infused error, not other types of error, for example coverage error.
• Margins of error vary by detailed or regional group, geography for detailed groups only, and table line. For example, household type (universe) and tenure (universe) tables have the following MOEs:
  • +/-3 households for detailed groups at nation and state.
  • +/-11 households for detailed groups at substate and AIANNH areas.
  • +/-50 households for regional groups at all geography levels.
• Counts for geographies that belong to multiple geography types are consistent across those types. Washington, D.C., is an example of these “geographic equivalents.” For tabulation purposes it is a state equivalent, county, and place. Although the disclosure avoidance system tabulates each geography type independently, the Census Bureau enforces consistency, so the data are demographically reasonable.
• Suppression addresses certain implausible scenarios imposed by disclosure avoidance. This includes the removal of negative values and “race alone” count when it is larger than the “race alone or in any combination” count.
• Aggregated data, such as geographic levels, household types, and detailed racial or ethnic groups, are less accurate than the specified MOEs. 
• Data are not always geographically consistent. The disclosure avoidance system generates a noise-infused count for each geography individually. This means that values from lower-level geographies may not sum to values from higher-level geographies.
• Data are not consistent between data products. The Detailed DHC-A and Detailed DHC-B data do not “match” comparable data from other 2020 Census data products, including the Redistricting File, Demographic and Housing Characteristics File (DHC), and the Demographic Profile.

Differences between Detailed DHC-A and Detailed DHC-B:

• The Detailed DHC-B uses a 90 percent confidence interval. This means that statistics in the Detailed DHC-B meet the established MOEs at least 90 percent of the time. The Detailed DHC-A uses a 95 percent confidence interval.
• The Detailed DHC-B has a larger relative error compared to the Detailed DHC-A. The Detailed DHC-A uses the same MOEs as the Detailed DHC-B, but the Detailed DHC-B universe is less than one-half the size of the Detailed DHC-A because there are fewer households than people. In addition, the Detailed DHC-B does not include the group quarters population (e.g., college residence halls, skilled-nursing facilities, military barracks).
• There are inconsistencies between household type and tenure tables. Both types of tables provide a total count of households (i.e., households and occupied housing units). The household count provided in the household type table may not match the occupied housing unit count in the tenure table. Alternatively, the total household count may be suppressed in one table but not the other.
• There are inconsistencies between the Detailed DHC-A and Detailed DHC-B. Since noise is infused separately to each, there might be more households for some detailed groups in the Detailed DHC-B than there are people in the Detailed DHC-A.
• Household type and tenure data may be provided for a group whose total population counts are suppressed in Detailed DHC-A. Noise-infused household type and tenure data are only produced when Detailed DHC-A noise-infused counts were produced. However, a group may receive household type and tenure data when a Detailed DHC-A total population count was suppressed.

The Census Bureau values your feedback on the Detailed DHC-B Proof of Concept. Please send feedback to 2020DAS@census.gov with the subject “2020 Census Detailed DHC-B.” To meet our 2020 Census Detailed DHC-B production release date of September 2024, all feedback must be received by February 28, 2024. If you do not want us to publish your feedback, or if you want us to remove your identifying information, please indicate that in the email.  

We would like to understand how you or your work will be impacted (positively, negatively, or not at all) if we release the final Detailed DHC-B using the current proposal outlined in the proof of concept. Please provide the level of geography, description of the use case(s), and implications should the data be released as reflected in this proof of concept. 

Please join us later today at 3:00 p.m. ET for a webinar to learn more about the Detailed DHC-B Proof of Concept and the next steps in finalizing the data product design and disclosure avoidance settings.

Log-In Details:

• Date: January 29, 2024
• Time: 3:00 – 4:00 p.m. ET. 
• WebEx link.
• Webinar/Access number (if needed):2831 084 8773
• Webinar password (if needed): Census#1

Today's Links:

• January 29 Webinar: 2020 Census Race and Ethnicity Data: The Detailed DHC-B Proof of Concept
• Detailed DHC-B Proof of Concept
• Detailed Summary Metrics
• Tenure Examples resource
• Graphics: Adaptive Design Process and Adaptive Design Summary
• Privacy-Loss Budget Allocation resource
• Frequently Asked Questions (FAQ) resource