XIV. Redistricting Data and Politically Sensitive Releases

The P.L. 94-171 redistricting file is generally the first detailed 2030 Census data product and one of the most politically consequential. It provides states with population counts for the small geographic areas used to draw congressional, legislative, and local election districts. For 2030, the file must be delivered to state officials by April 1, 2031.

In 2020, the redistricting file included population counts by race, Hispanic origin, and voting age, along with information about group quarters and housing occupancy. These data were available for blocks, tracts, counties, voting districts, and other geographies.

The 2030 file will be shaped by several decisions made elsewhere in the census. Questionnaire decisions will determine which racial and ethnic information Census collects. Disclosure avoidance will determine how much of that information can be published for small areas. Geographic programs will determine the blocks and voting districts for which the data are tabulated. Release procedures will determine whether state officials receive the information before the public.

These are not separate technical questions. Together, they determine whether redistricting data are detailed enough to identify communities, released fairly enough to earn public trust, and stable enough to support a rapidly changing legal and political environment.

A. Race and Ethnicity Content

Race and ethnicity are central to the redistricting file because the data are used to evaluate whether district plans provide equal electoral opportunities and comply with federal voting-rights law. Voting-age information is cross-tabulated with race and Hispanic origin so users can examine the population eligible to vote within proposed districts.

The content of the file begins with the questionnaire. If Census removes a category or collects less detailed information, disclosure avoidance cannot restore what was never collected. If Census preserves or expands the available categories, the redistricting file must decide how those responses will be tabulated.

Disclosure avoidance creates a second decision point. Census may collect a person’s race and ethnicity accurately but later combine categories, suppress values, or publish the information only for larger areas. The person remains somewhere in the total population, but the community may no longer be visible in the local data used for redistricting.

These pressures increase when several dimensions are combined. A table showing race, ethnicity, and voting age for a small block can contain many cells with only a few people. More inclusive categories may improve the underlying data while also increasing the number of small cells that require protection.

Census should therefore design the questionnaire, redistricting tables, and disclosure avoidance system together. It should publish early proposals showing:

  • which race and ethnicity categories will appear in the file;

  • how multiple responses will be tabulated;

  • whether any categories will be combined for small geographies;

  • which tables will include voting-age information; and

  • how disclosure avoidance will affect the ability to identify smaller populations.

The redistricting file should also clearly distinguish between information that Census did not collect and information that it collected but cannot publish at a particular level of detail. Those outcomes may look similar in the final table, but they reflect very different policy choices.

Central question

Will the 2030 redistricting file preserve enough race, ethnicity, and voting-age detail to identify communities at the geographic scale where districts are actually drawn?

B. Geographic Detail

Redistricting depends on small geographic building blocks. States and local governments assemble census blocks into congressional districts, legislative districts, city wards, school-board districts, and other election areas. Blocks allow map drawers to move a boundary without being forced to divide a larger census tract or municipality.

P.L. 94-171 requires Census to give states an opportunity to identify the small areas for which they need population counts. The 2030 Redistricting Data Program allows states to suggest block boundaries and submit voting-district geography for use in the final tabulations.

The disclosure avoidance DAO may be difficult to reconcile with this geographic structure. Blocks often contain few people, and detailed race and voting-age tables create many small cells. If Census must protect those cells primarily through coarsening and suppression, it may have to combine blocks, reduce demographic detail, or withhold portions of the table.

That does not necessarily mean all block data will disappear. Census may be able to preserve exact or highly accurate total-population counts while reducing the associated demographic detail. But a block total alone is not enough for many redistricting and civil rights uses. Users need to know where racial and ethnic populations live and how those populations are distributed among voting-age groups.

Geographic aggregation also creates practical problems. Census tracts and block groups were designed for statistical purposes and do not necessarily align with neighborhoods, precincts, tribal communities, or proposed district boundaries. Combining blocks in advance may force map drawers to use geographic units that cross the very communities they are trying to identify.

Census should publish separate commitments for:

  • total population by block;

  • voting-age population by block;

  • broad race and ethnicity data;

  • detailed or multiracial tabulations;

  • group quarters populations; and

  • state-submitted voting districts.

It should also test whether the remaining geography supports local redistricting, not only congressional and state legislative maps. A system may provide enough information to divide a state into several large districts while failing for city councils, county boards, or school districts whose populations are much smaller.

Credible risk

Census technically provides block-level totals but removes enough demographic detail that users cannot evaluate whether district boundaries divide or dilute small racial and ethnic communities.

C. Pre-Release Access and Embargoes

The Operational Plan proposes giving official state recipients embargoed access to the redistricting data before the public release. Census had planned a similar process for 2020 but suspended it after pandemic-related delays. The final 2020 legacy-format files were released to state officials and the public simultaneously.

Historically, official recipients have included governors and legislative leaders. For 2030, each state also appoints a nonpartisan liaison who serves as Census’s point of contact throughout the Redistricting Data Program. The current plan does not yet identify exactly which individuals will receive embargoed access or how long the embargo will last.

A short embargo can have a legitimate operational purpose. State officials can confirm that files were received, verify that they open correctly, and alert Census to an obvious delivery failure before a public announcement. This may be especially useful when the files are large or technically difficult to process.

But redistricting data are not an ordinary statistical release. Even a brief advantage could allow officials to begin analyzing districts, preparing public narratives, consulting political advisers, or developing litigation strategy before opposition parties, civil rights organizations, journalists, and the public can see the same information.

The concern is greatest where the official recipient is also directly involved in drawing or approving maps. A statutory responsibility to receive the file does not necessarily require an extended period in which the recipient may use the data privately.

If Census retains an embargo, it should establish strict and public rules:

  • access should be limited to specifically named officials and technical staff;

  • the embargo should be the same length in every state;

  • data should be provided through secure, individually authenticated accounts;

  • downloading, sharing, and access should be logged;

  • recipients should be prohibited from using the data for map drawing, public messaging, or partisan planning before release; and

  • violations should carry clear consequences.

Security in this context serves a somewhat different purpose from Title 13 confidentiality. The embargoed file will already have undergone disclosure avoidance and will soon become public. The principal concerns are premature disclosure, unequal political advantage, and confidence in the integrity of the release.

Simultaneous public release remains the clearest equal-access model. If Census concludes that an embargo is operationally necessary, it should be limited to the shortest period needed to verify delivery and technical usability. Census should publicly identify the categories of recipients, the terms they accepted, and the time at which access was granted.

The Bureau should also avoid rolling releases that give some states access days or weeks before others without an unavoidable operational reason. A single national release would reduce concerns that the timing itself favors particular jurisdictions or political actors.

Open design question

Who will receive embargoed 2030 redistricting data, for how long, and what may they do with the information before it becomes public?

Advocacy priority

Prefer simultaneous release. Any embargo should be brief, uniform, technically justified, and governed by enforceable restrictions on use and sharing.

E. Count Question Resolution

Count Question Resolution, or CQR, provides a limited process for tribal, state, and local governments to ask Census to review certain errors after the census results are released. Despite its name, it is not a general appeal of the census count and does not allow communities to reopen the enumeration.

For 2030, the Operational Plan says CQR will review possible boundary errors and errors involving the number or geographic placement of living quarters, including housing units and group quarters. The review is based on information Census already collected and processed. It does not conduct new interviews or a recount.

CQR may be able to correct situations such as:

  • housing units assigned to the wrong census block;

  • a group quarters facility placed on the wrong side of a boundary;

  • living quarters omitted because of a processing error; or

  • an incorrect governmental boundary used to allocate population.

CQR generally cannot correct:

  • a claim that residents were missed because fieldwork was inadequate;

  • an undercount that would require new data collection;

  • incorrect race, ethnicity, age, or other characteristics;

  • dissatisfaction with disclosure avoidance;

  • disagreement with Census residence rules; or

  • an error in a district map drawn by a state.

The process is also available to governmental units, not directly to individual residents or community organizations. Advocates must generally work through an eligible tribal, state, or local government to submit a case.

Most importantly, a successful CQR challenge does not rewrite the official decennial products. Under the current plan, corrected counts will not replace the apportionment totals, P.L. 94-171 redistricting file, or other 2030 Census data products. The corrections will instead inform later population estimates and may be provided to the affected government for planning and funding purposes.

That limitation can be frustrating when Census confirms a real error after districts have already been drawn. But changing the official redistricting file jurisdiction by jurisdiction could also create serious consistency and timing problems. Census should therefore invest heavily in pre-release geographic review, count review, and quality control rather than relying on CQR to repair the data later.

Census should publish the 2030 CQR rules well before the census, including eligible errors, evidentiary requirements, deadlines, and expected remedies. It should also maintain a prompt errata process for clear production errors that do not require a full CQR case.

Important limitation

CQR can correct certain boundary and living-quarters processing errors for future statistical uses. It cannot conduct a recount or generally revise the official redistricting data.

Why advocates should care

Communities should understand CQR before relying on it as a remedy. Many of the most serious coverage and data-quality concerns must be prevented or challenged earlier because they cannot be corrected after the redistricting file is released.

Where to Look: See the 2030 Census Operational Plan, especially section 3.4.2, Redistricting Data Program, pp. 61–63, for the redistricting-data process, proposed embargoed access for official state recipients, geographic updates, and disclosure avoidance. Section 3.4.3, Count Question Resolution, pp. 63–64, describes the limited post-census process for reviewing boundary and living-quarters errors.

The Census Bureau’s 2030 Redistricting Data Program webpage is the best source for the evolving schedule, state liaison process, voting-district and block-boundary programs, and eventual P.L. 94-171 specifications. The 2020 P.L. 94-171 Summary File provides the most useful baseline for understanding the race, ethnicity, voting-age, group quarters, and geographic detail traditionally available for redistricting.

For post-release corrections, see the 2020 Count Question Resolution materials and operational assessment, which show what kinds of errors CQR can review and how corrected counts are handled. For 2030, advocates should also watch future Operational Plan updates for the final embargo rules, disclosure avoidance decisions, race and ethnicity content, and any changes to the geographic detail available in the redistricting file.

****