WASHINGTON — A new court filing has cast fresh doubt on the accuracy of the list at the centre of the Justice Department’s pre-election fight with California, three days before the midterm elections. An analysis by two researchers from the University of California, Los Angeles Voting Rights Project, submitted to a federal court on Saturday, found that 8,880 of the 20,930 names the department flagged as likely noncitizen voters — about 42 per cent — could not be matched to anyone on California’s current voter file. The researchers also identified at least 22 people on the list who are confirmed United States citizens, with 66 more rated as near-certain matches to naturalisation records, and estimated that nearly 7,500 naturalised citizens were highly likely to have been flagged incorrectly under the government’s matching methods.
The filing, submitted by the Coalition for Humane Immigrant Rights, known as CHIRLA, is the first outside analysis of the full federal list, which the department had kept closely held until a judge ordered its disclosure. It does not settle the underlying court cases, and it is not a judicial finding. But its timing — filed on Saturday, with voting already under way and Election Day on Tuesday, November 3 — has made it one of the most closely watched documents of the final weekend, because the list itself was the basis for the department’s demand that California election officials separate, or sequester, ballots tied to the flagged names during the count.
The Justice Department has defended its list and its broader voter-enforcement campaign. Federal officials say analysts compared federal immigration data with California voter records and identified more than 20,000 people they believed were likely noncitizens who had received vote-by-mail ballots. A department spokeswoman has said that every ballot cast by an ineligible voter cancels the vote of an American citizen. Bill Essayli, head of the US Attorney’s office in Los Angeles, dismissed the new analysis on Saturday, describing it as a ‘bogus report’ produced by what he called an activist immigrant-rights group. That clash — a plaintiffs’ expert report on one side, a flat official rejection on the other — is now part of the court record, and judges, not press statements, will decide what weight it carries.
How the UCLA researchers tested the list
The researchers compared names, dates of birth and addresses on the federal list against California’s voter file and citizenship records. Their work was possible only after Chief Judge James Boasberg ordered the Department of Homeland Security to hand the list to CHIRLA’s lawyers in their lawsuit challenging the administration’s voter-enforcement campaign. Until that order on Friday, the list had been described by officials but not disclosed in full to the plaintiffs challenging it, which meant its accuracy could be debated in general terms but not tested name by name.
The testing itself ran into problems in the data, the researchers reported. The federal list lacked a state voter-identification number for each entry — the key data point election officials use to confirm matches and detect duplicates — and it contained duplicates and incomplete or inconsistent data, making one-to-one verification more difficult. Those gaps matter for two reasons. First, they help explain why a name might fail to match even if the person behind it is a real voter. Second, they are themselves part of the plaintiffs’ argument: a list that election officials were asked to use, in the middle of an election, to separate ballots should be precise enough to use safely, and the researchers say this one was not.
The headline figures need careful reading, and the researchers themselves caution against over-reading them. The 8,880 unmatched names are names that could not be matched to California’s current voter file. That does not by itself prove that every one of those people is a citizen, or that none of them ever appeared on an older or inactive registration. Some unmatched names may reflect old, cancelled or inactive registrations, or data errors on either side. What the mismatch does establish, the filing argues, is the scale of the problem in a list that officials asked counties to act on: more than four in ten names could not be tied to a current voter at all. The 22 exact citizen matches are firmer ground — people whose naturalisation records confirm citizenship — and the 66 near-certain matches and the estimate of nearly 7,500 naturalised citizens highly likely to have been misflagged rest on the researchers’ matching methods, which the government disputes.
What the California voter list accuracy findings do and do not prove
The distinction between a plaintiffs’ expert report and a court ruling is central to this story, and it is where both sides’ claims need qualifying. The UCLA analysis is evidence submitted by one side in litigation, prepared by researchers working for the plaintiffs. It has not been adopted by a judge as a finding of fact, and no ruling issued on Saturday ordered ballots rejected, segregated or counted differently because of the new report. At the same time, the report’s central findings were described by the Associated Press, which reviewed the filing, in the same terms as other outlets: 8,880 names unmatched, at least 22 confirmed citizens. The administration’s response so far has been to attack the messenger rather than to publish a competing name-by-name analysis, and no response from the department or from California Secretary of State Shirley Weber to the new analysis itself was reported in the coverage reviewed for this article.
The filing also lands against a backdrop of repeated court defeats for the department’s ballot-separation demand. The department sent California its list on October 12 and asked the state to sequester the corresponding ballots. California refused, and the department sued on October 15. A federal district judge denied emergency relief, and on October 30 a three-judge panel of the US Court of Appeals for the Ninth Circuit rejected the administration’s renewed emergency request in a 2-1 decision, leaving the denial in place. Separately, US District Judge Trina Thompson issued a temporary restraining order, running through November 16, barring the department from threatening, intimidating or coercing state or local election officials into halting the processing of a registered voter’s ballot solely because the administration alleges that person cannot vote. The administration appealed that order on Saturday, according to NBC News reporting. Readers can follow the earlier court chapter in our coverage of the California ballots Supreme Court showdown and the earlier California ballots ruling blocking the segregation demand.
Federal law bars noncitizens from voting in federal elections, and that point is not in dispute between the parties. What is disputed is how reliably the government can identify noncitizen voters from administrative data, and what election officials should be asked to do on the basis of such a list while an election is already under way. California and Los Angeles County officials have challenged the department’s methodology throughout, and a preliminary county review found that many records involved inactive or cancelled registrations. The UCLA filing adds a statewide, name-level dimension to that challenge: it is not only that the methodology is contested in theory, the plaintiffs argue, but that its output, when checked, fails to match the voter file in more than four cases in ten.
Why the filing matters before Tuesday — and what happens next
The practical stakes are narrower than the rhetoric on either side might suggest, and they are worth stating plainly. No court has ordered California to separate the disputed ballots, and the new filing does not itself change how any ballot will be counted on Tuesday. California’s results are not certified until December 11, and the litigation over the department’s demands — including the Thompson order, the Ninth Circuit proceedings and the CHIRLA case itself — will continue after Election Day. The filing’s immediate effect is therefore on the argument, not the count: it gives judges and the public a concrete, checkable set of numbers about a list that had previously been debated largely in the abstract.
That is also why the administration’s next move matters. If the department produces its own matching analysis, or discloses more about how the list was built, the debate can move from competing characterisations to comparable data. If it does not, the UCLA numbers — unmatched names, confirmed citizens, likely misflagged naturalised citizens — will remain the only name-level test in the public record, disputed but unanswered in kind. Either way, the episode has already exposed a structural tension that will outlast this election: federal agencies hold immigration and citizenship data, states hold voter rolls, and joining the two datasets without voter-identification numbers, with duplicates and inconsistent fields, produces errors that fall hardest on naturalised citizens, whose records are the most likely to be mismatched across systems.
For voters in California, the message from election officials has been consistent: eligible voters should vote as normal, and no voter should assume their ballot has been set aside because of this dispute, since no court has ordered such a step. For the rest of the country, the California filing is a preview of a question that could recur wherever federal voter-enforcement efforts meet state-run elections: how accurate must a government list be before officials are asked to act on it, and who gets to test it before it is used? The courts will answer that question case by case. The UCLA analysis has at least made the question concrete, with numbers that both sides will now have to reckon with.
Sources
USN reviewed reporting from multiple outlets for this story. Read more from Al Jazeera, Times Now, WLTReport.