The Preliminary Number Problem
Scott Presler's viral post says 250,000 non-citizens are on the voter rolls in four states. History suggests that number is a starting point for an investigation, not a finding — and it usually shrinks.
The figure is real, in the sense that a federal agency really did put it out. On July 17, 2026, DHS Secretary Markwayne Mullin announced that a preliminary review had flagged roughly a quarter million potential non-citizen registrations across California, New Jersey, Nevada, and Pennsylvania. Presler's post repeats that number accurately. What it doesn't repeat is the word DHS itself used to describe it: preliminary.
What DHS actually said
| State | Flagged registrations |
|---|---|
| California | 190,832 |
| New Jersey | 35,152 |
| Nevada | 15,903 |
| Pennsylvania | 14,576 |
Source: DHS letters to state election officials, July 17, 2026 — reported by Fox News, Newsweek, and DHS.gov.
The method behind those numbers: DHS compared commercial databases and federal immigration records against public voter files, looking for name, date-of-birth, and identifier matches. That is not a citizenship check performed on individual people. It's a bulk pattern match — the same basic method that has already been tried, at scale, in at least three other states. Every time, the number the public first heard was much larger than the number that survived scrutiny.
The same story, three times already
Gov. Greg Abbott announced 6,500 noncitizens had been cut from the rolls. A joint investigation by ProPublica, the Texas Tribune, and Votebeat found the state's own report to the governor identified only 581 as noncitizens — the rest were people who simply hadn't answered a citizenship-confirmation letter, which is not the same thing as not being a citizen.
Secretary of State Wes Allen instructed all 67 counties to inactivate 3,251 voters flagged by DHS identification numbers. Court filings later showed the state itself acknowledged at least 2,074 of them — well over half — were legally entitled to vote all along. A federal judge halted the program.
A routine quarterly DMV cross-check usually surfaces about 200 names. This round surfaced 3,500. The state's elections director told legislators the likely cause was old records — people who naturalized after getting a license and never went back to update it — not a wave of noncitizen registrants.
What's confirmed, and what isn't
- DHS sent letters to CA, NJ, NV, and PA citing these exact figures.
- The method was a database/records match, not individual verification.
- This same match-then-flag approach has, in every prior public test (TX, AL, AK), produced numbers that shrank sharply on review.
- Nevada's and Pennsylvania's Secretaries of State have already disputed the figures on the record.
- That 250,000 people are actually non-citizens.
- That any of the flagged registrants ever cast a ballot.
- That the four-state number will hold up the way the Texas and Alabama numbers didn't.
- Any independent audit of the DHS match — none has been published.
The honest takeaway
None of this means zero non-citizens are registered anywhere — Alabama, after full review, did confirm a small number, and Texas confirmed 581 over three years. Errors on voter rolls are real and worth fixing. What the record doesn't support is treating a first-pass database match as a finding. Nevada's Secretary of State called the DHS numbers "wildly speculative at best" and said the department "hasn't shared anything that backs it up." Pennsylvania's said "all evidence has shown that noncitizen voting is extremely rare."
Presler's post isn't fabricating a number — DHS really did publish it. But presenting it as settled fact, with no mention that it's a preliminary estimate built on a method with a documented track record of over-flagging citizens, is exactly the gap between "a real number was announced" and "250,000 non-citizens are voting." Given what happened the last three times this exact type of claim was made, the safer bet is that this one shrinks too.
Posts
No comments:
Post a Comment