New: DHC Demonstration Data for Housing Tables
Next Steps in DHC Development
March 29, 2022: Today the Census Bureau released 2010 housing demonstration data for the proposed 2020 Census Demographic and Housing Characteristics File (DHC).
The demonstration data apply the current iteration of the disclosure avoidance system (DAS) to 2010 Census data. This allows a side-by-side comparison of 2010 Census data: one protected with the DAS and one (the published tables) protected with the legacy swapping methods.
How to Submit Feedback
We ask data users to submit feedback by April 29, 2022, on both the person tables released March 16 (tables | newsletter) and the housing tables released today. Together the person and housing tables comprise the first of two rounds of demonstration data and associated comment periods.
We would like to hear from you about how you use these data and how those uses are impacted if we release the final data product using the current data protection settings. Please provide the table, level of geography, description of the use case, and implications should the data be released as reflected in this demonstration data release.
Please send feedback to 2020DAS@census.gov using the subject “2020 Census Data Products.” It is vital that we receive your comments by the April 29 deadline to ensure that we produce the final data product in a timely manner.
It’s possible that we might choose to publish detailed summaries of the comments we receive. If so, we will make every effort to remove identifying information, such as names of individuals and organizations.
About This Release
-
Content. Today’s release includes the HOUSING demonstration data from the suite of proposed DHC tables. Note that these housing data are also referred to as “unit” data. Housing tables include content on tenure, vacancy, household type, family type, and characteristics of the householder. Person tables (released March 16) include content on sex, age, Hispanic origin, race, relationship to householder, and group quarters. Some housing and person tables are repeated by race and ethnicity. You can see which tables are housing tables in the March 16, 2022, technical documentation as well as the updated 2020 Census Data Product Planning Crosswalk (see the “DHC Crosswalk” tab, column H).
-
Detailed Summary Metrics. Detailed Summary Metrics for the housing tables were released on March 16 along with the metrics for person tables. The metrics are designed to help identify accuracy, potential bias, and outliers. Keep in mind that the published 2010 data include error due to “swapping,” the disclosure avoidance method used at the time. The demonstration data apply differentially private protections to the unswapped, confidential 2010 data. Neither the published 2010 data nor the 2010 demonstration data necessarily reflect the values of the true, unprotected data. So, a mean absolute error of five households in the metrics, for example, reflects the differences between the published (swapped) data and the differentially private data.
-
Progress on the Housing Tables Since the DHC DAS Demonstration Data Released in October 2019 and May 2020. Although we released DHC demonstration data based on the 2010 Census in October 2019 and May 2020, those releases were based on early iterations of the DAS and included only a subset of proposed DHC tables. For areas that were included in previous releases, some have seen significant improvements in accuracy. The graphic below shows that error has decreased for every household type, but particularly for family households with a female householder and no spouse present.
-
Areas We’re Still Working to Improve in the Housing Tables. DAS development is an ongoing, iterative process. We’ve identified areas for improvement in the housing tables and are actively working to address them prior to release of the second round, including:
- Additional accuracy for homeownership rate for householders under 35
- Additional accuracy for homeownership rate by race and Hispanic origin
- Reduced positive bias for households with own children under age 6
- Additional accuracy for same-sex married and unmarried couple households
- Reduced negative bias for one- and two-person households and positive bias for households with seven or more persons
We are continuing to review all tables to identify additional areas for improvement.
-
Updated Technical Documentation. The technical documentation for person and housing tables was released on March 16. For users who analyze the demonstration data via the Summary File format, there was an error in the housing table segments. We updated the technical documentation to fix that error along with a few additional user-friendly updates.
-
Table Format Only; No Microdata. The data are in table format, not microdata format. While microdata are used to generate the tabular data, they include details at geographic levels we are not planning to publish for certain tables. Forgoing publication of a privacy-protected microdata file (PPMF) for the demonstration data allows us to achieve greater accuracy for the published tables for the same level of confidentiality protection.
-
Watch the Webinar for More Background. Census Bureau staff described DHC development and answered questions in a webinar on March 22. Visit the webinar page for the recording, transcript, and slides.
How the DHC Disclosure Avoidance System (DAS) Development and Feedback Rounds Work
In addition to this release, we will release one more round of demonstration data that incorporates feedback from the public. Although previous demonstration data had limited DHC content (October 2019 and May 2020), we are referring to the current and future releases as rounds one and two for simplicity.
Here’s how the DAS development and feedback cycle will work. Note that we are finalizing target dates for each of these steps and will release them shortly.
-
Round One Feedback and Analysis. Census Bureau subject matter experts will review and analyze all comments on this first round of demonstration data (both the person and housing files) submitted by April 29 via our 2020DAS@census.gov.
-
System Tuning and Impact Analysis. Based on feedback submitted during the first round, our subject matter experts will address the data quality issues that warrant further system tuning. Those decisions will factor in the potential costs and benefits of changes to the DAS in terms of their impacts on confidentiality and accuracy. Our science team will then conduct a series of experiments to optimize these feedback-based adjustments. We will compare the outcomes of those experiments against metrics developed to measure fitness-for-use across variables and geographies. (Detailed Summary Metrics | Metrics Overview.)
-
Demonstration Data Content and Settings Selection for Round Two. The Data Stewardship Executive Policymaking Committee (DSEP), composed exclusively of senior career officials, will select the content and settings for the second round of DHC demonstration data. Their decision will follow discussion of analysis and recommendations submitted jointly from leadership of both the Research and Methodology Directorate and the Demographic Directorate.
-
Round Two Demonstration Data Feedback. The second round of demonstration data will reflect changes adopted from first-round comments. It will also reflect some table changes (additional tables and geographical granularity) based on public comments on the 2020 Census Data Product Planning Crosswalk released September 16, 2021 (updated March 16, 2022). The final scope and granularity of those proposed changes are dependent on ongoing analysis. You will have 30 days to sumit your feedback regarding the data’s fitness-for-use for your data needs. By the time we issue the second round of demonstration data, the DAS will be in near-final form; the tuning possible based on comments at that point will be more limited than for the first-round demonstration product. In addition, there will be a second round of tuning and impact analysis which will result in recommendations for the final content and settings for the DHC.
-
Final DHC Settings Selection. The DSEP will determine the final content and settings for the DHC, based on public feedback and the joint recommendations of the Research and Methodology Directorate and the Demographic Directorate.
-
System Programming, Quality Assurance Analysis, and Final DHC Data Product Release. After the content and settings are finalized, the final system programming and quality assurance begins. This is an extensive process and we’ll provide more information about it, as well as expected timing, duration, and the target release date, in the coming months.
-
Release of Demonstration Data Reflecting Final Production Settings. As with the 2020 Census P.L. 94-171 Redistricting Data release, we will release a final DHC demonstration data product reflecting the chosen content and settings used for production. The production-settings release is intended for fitness-for-use analysis purposes only, not for further feedback.
Recently Posted:
Useful Links:
Was this forwarded to you?
Sign up to receive your own copy!
|
