Of 20,930 people on a Justice Department list of alleged noncitizen voters in California, 8,880 — 42 percent — could not be matched to anyone on the state’s current voter file, according to an outside analysis disclosed Saturday, three days before the midterm election.
The review, produced by UCLA Voting Rights Project researchers for plaintiffs challenging the administration’s voter-enforcement campaign, also found at least 22 exact matches to people whose naturalization records confirm they are U.S. citizens. The figures appear in a new court filing and accompanying expert report; an independent AP review reported the same central findings.
The analysis is not a judicial finding and does not establish that every name on the federal list is erroneous. But it is the first broad outside test of a dataset the Justice Department has cited while seeking to have California flag and segregate ballots associated with the listed people. Its scale and the confirmed citizens it identified materially sharpen a legal fight over who may vote, what evidence is sufficient to challenge eligibility and how far federal officials may press state election workers immediately before voting ends.
How the federal list was tested
The Justice Department assembled the list by comparing federal immigration data with California voter records and said the 20,930 people were “likely noncitizens.” In its lawsuit, the department argued that California should identify affected ballots and keep them separate until eligibility could be resolved. The state has disputed both the federal methodology and the legality of the requested intervention, while a federal appeals court this week rejected a separate administration effort to segregate roughly 20,000 California mail ballots in an earlier ruling.
The plaintiffs’ researchers received the underlying list only after U.S. District Judge James Boasberg ordered the government to disclose it. They compared names, dates of birth and addresses against California’s voter file and citizenship records. According to the filing, the list lacked a state voter-identification number for each entry, contained duplicates and included incomplete or inconsistent data, making one-to-one verification more difficult. The reported findings include 66 additional people whom the plaintiffs say are highly likely to be citizens based on near-exact matches to naturalization records, beyond the 22 exact matches.
Using the government’s own matching assumptions, the AP calculated that nearly 7,500 naturalized citizens were highly likely to have been flagged incorrectly. That estimate is broader than the 22 confirmed exact matches and depends on probabilistic matching, so it should not be read as a verified count of erroneous entries. It does, however, show why election officials and the court cannot treat the list itself as a final determination of citizenship or voting eligibility.
Earlier county checks raised similar concerns
Los Angeles County had already tested a portion of the federal material before Saturday’s statewide analysis. Of 7,035 county records it examined, only 42.4 percent were active registrations; 42.9 percent had been canceled and 4.6 percent were inactive. In a closer sample, 25.8 percent of the named people had a U.S. birthplace in county records, and 71 percent had never voted, according to the county data summarized in a county sample and cited in the plaintiffs’ filing.
Those categories do not by themselves prove citizenship. A U.S. birthplace generally indicates citizenship, but local voter files can be incomplete, and canceled or inactive registrations are not the same as active illegal voting. The county results nevertheless matter because they independently undercut the idea that every name represents a current, active, noncitizen voter. They also reinforce a basic distinction blurred by the federal count: a data match is a lead for investigation, not proof that a person is ineligible or has cast a ballot.
The Justice Department stands by its campaign
The Justice Department has defended the list and its demand for action. Assistant Attorney General Harmeet Dhillon said the government would not apologize for protecting citizens’ votes and argued that every ineligible ballot dilutes a lawful one, according to AP reporting. The department’s public account says its matching process identified likely noncitizens, not conclusively proven ones, and frames ballot segregation as a temporary safeguard while eligibility is checked.
The dispute is therefore partly about legal authority and partly about the reliability threshold required before election officials burden a voter. In a related case, Kollar-Kotelly barred federal officials from threatening California election workers with prosecution or financial penalties for declining to follow federal demands; that threats order remains distinct from the factual question of who appears on the list. Saturday’s evidence does not resolve either case, but it gives the court the first large-scale measurement of the list’s apparent mismatch rate.
What changes before Election Day
The immediate practical effect is to make broad action against all 20,930 entries harder to justify without individual verification. California election officials remain responsible for determining eligibility and counting ballots under state and federal law, while the Justice Department can continue presenting evidence in court. No ruling issued Saturday ordered ballots rejected, segregated or counted differently because of the new report.
The next consequential evidence will be the government’s response to the matching analysis and any court determination about the list’s reliability. For now, the record establishes that a substantial portion of the federal list does not correspond to the current statewide roll and that confirmed citizens appear among the names. It does not establish the final status of every person listed, nor does it measure how many, if any, ineligible ballots were cast.
