The Road Ahead: Upcoming Disclosure Avoidance System Milestones

Registered United States Census Bureau Logo
Privacy lock

The Road Ahead: Upcoming Disclosure Avoidance System Milestones

Recently, the Census Bureau announced the official schedule for delivery of the redistricting (P.L. 94-171) data product. With release by September 30, 2021, now confirmed, the Disclosure Avoidance System (DAS) team has finalized the schedule for the stakeholder engagement, decision-making, and system improvements necessary to prepare the DAS to produce the files. While minor adjustments to these dates may still occur, we want to inform our data users of what they should expect over the next several months, and how they can continue to provide feedback and input to this process.

According to this schedule, the DAS will begin the final production run for the 2020 Census redistricting data product in late June 2021. Once completed, the privacy-protected redistricting files (known internally as the Microdata Detail Files) will undergo extensive review and quality assurance by subject-matter experts in the Census Bureau’s Population Division before proceeding to tabulation and release.

Prior to the late June production run, the Census Bureau’s Data Stewardship Executive Policy (DSEP) Committee will set the final privacy-loss budget (PLB) and production parameters for the DAS run of the redistricting data product. We currently expect these DSEP decisions to be finalized the first week of June 2021.

New Demonstration Data with Higher PLB

Over the 16 months since our first release of DAS demonstration data products using the 2010 Census, we have benefited greatly from engagement with and input from our data users. With each subsequent release of demonstration data (in May 2020, September 2020, and November 2020) we have received extensive actionable feedback from the data user community, feedback that has informed ongoing DAS system improvements and design changes.

Throughout this process, however, we maintained the conservative PLB set for the initial demonstration data product. While we recognize that this decision to hold the PLB constant across the demonstration runs meant that the resulting data would have substantially more noise (error) than should be expected in the final 2020 Census data products, holding the PLB constant enabled us and our data users to home in on the elements of the algorithm that were causing systemic distortions that needed to be addressed. We acknowledge that this has unfortunately led some of our data users to expect comparable amounts of noise in the final 2020 Census data.

In our last newsletter, we announced our intention to produce another set of Privacy-Protected Microdata Files (PPMFs) and Detailed Summary Metrics for our data users to evaluate before DSEP sets the final PLB and DAS parameters for the 2020 Census redistricting data product. These demonstration data will feature a higher PLB and system parameter optimization informed by the hundreds of full-scale DAS experimental runs we have been performing over the last several months. (Stay tuned for more information about what we’ve learned from these experiments in our next newsletter.)

The resulting data will more closely approximate the expected accuracy and fitness-for-use of the final 2020 Census redistricting data product. More importantly, these demonstration data will enable our data users to provide critical fitness-for-use analyses that will inform DSEP’s decision-making.

After the June DSEP decision, and before the September redistricting data release, we will release a final set of PPMFs reflecting the chosen PLB and system parameters.

Ensuring Data User Evaluation Time

We are planning to release the next PPMFs and Detailed Summary Metrics no later than April 30, 2021, thus providing our data users with over four weeks to perform their analyses and submit feedback and recommendations prior to DSEP’s June decision. In the unlikely event that operational delays impact delivery of the new PPMFs by the April 30 deadline, we will delay the DSEP decision meeting to ensure that our data users have a full four-week period to review the demonstration data and provide feedback to inform these critical decisions.

 

Recap: 2021 Key Dates, Redistricting (P.L. 94-171) Data Product

By April 30:

  • Census Bureau releases new PPMFs and Detailed Summary Metrics

By late May:                

  • Data users submit feedback

Early June:                   

  • DSEP makes final determination of PLB, system parameters based on data user feedback

Late June:                    

  • Final data production run and quality control analysis begins

September:                 

  • Census Bureau releases PPMF and Detailed Summary Metrics from applying the production version of the DAS to the 2010 Census data
  • Census Bureau releases production code base for P.L. 94-171 redistricting summary data file and related technical papers

By September 30:         

  • Redistricting data release

DAS Webinar Series

We also acknowledge that many of our data users have asked for more information about how the DAS operates, what improvements we have made over the past several months, and what their potential impact on the data may be for various use cases. To provide greater transparency into these (and other) issues, we are happy to announce our intention to host a series of webinars this spring and summer. If you have suggestions for topics you would like to see covered in this webinar series, please submit them to 2020DAS@census.gov. We will announce the schedule and details for this webinar series in a future newsletter.

Looking Ahead to the Demographic and Housing Characteristics Data

As we have previously discussed, changes to the operational schedule required us to focus our efforts on preparing the DAS for production of the redistricting data product. With our optimization experiments for the redistricting data nearing completion, we are beginning to turn our attention back to optimizing the DAS to ensure fitness-for-use for the next scheduled 2020 Census data products, the Demographic Profiles and Demographic and Housing Characteristics (DHC) files.

We will provide more information regarding these efforts in subsequent newsletters, along with information about forthcoming DHC-specific demonstration data that we intend to release.

Was this forwarded to you?

Sign up to receive your own copy!

Sign Up!


Useful Links:


Have Suggestions?

Do you have specific questions you'd like us to answer in this newsletter or topics you'd like discussed? Send us an email at 2020DAS@census.gov and let us know!

Contact Us

About Disclosure Avoidance Modernization

The Census Bureau is protecting 2020 Census data products with a powerful new cryptography-based disclosure avoidance system known as “differential privacy.”  We are committed to producing 2020 Census data products that are of the same high quality you've come to expect while protecting respondent confidentiality from emerging privacy threats in today's digital world. 

 

Share This