Election denial has lately come to be viewed as a feature of the political right, reflected by the lawsuits, conspiratorial documentaries, and “Stop the Steal” protests that followed Donald Trump’s loss in the 2020 presidential election. But in the months since 2024, a similar—albeit much quieter—form of election denial has emerged in parts of the progressive left.
These theories range from claims that Elon Musk used Starlink satellites to hack the election to a the quasi-mystical TikTok subculture known as the “4 A.M. Club,” whose members believe the timeline glitched and Kamala Harris won in a parallel reality. But the most prominent claims have been rooted in data-heavy spreadsheets and statistical jargon.
One of the most popular of these theories suggests that a 2024 National Security Agency audit confirmed that Kamala Harris won the election, a claim which gained notoriety after it appeared in This Will Hold, an anonymously published Substack. The post alleges that one of the audit’s supposed participants, an ex-CIA officer named Adam Zarnowski, possessed insider information about a global cabal of corrupt actors, international criminals, foreign operatives, billionaires, and political insiders who conspired together to manipulate the election’s outcome.
As The Atlantic recently reported, there is no independent verification of Zarnowski’s background beyond his own claims. A LinkedIn profile describes him as a “former CIA paramilitary operations officer” but provides no evidence that he is an expert in election security or statistics. Snopes has been unable to “independently verify Zarnowski’s employment with the CIA or his alleged involvement in [the] NSA audit.”
The Election Truth Alliance (ETA), a self-described nonpartisan watchdog group, has used statistical models to push claims that Harris won the election. In Rockland County, New York, for example, Harris received fewer votes for president than incumbent Sen. Kirsten Gillibrand (D–N.Y.) did for Senate. The ETA suggests that possible election tampering can be inferred from this discrepancy.
But Charles Stewart, a political scientist at the Massachusetts Institute of Technology, points out that this apparent discrepancy isn’t unusual and can easily be explained. Stewart attributes Harris’ weaker performance to her unpopularity among the county’s Orthodox Jewish voters relative to Gillibrand, as well as the broader trend of voters skipping races or voting split-ticket.
The organization’s claims go further. In a recent interview with the progressive commentator David Pakman, the ETA’s Nathan Taylor claimed that vote patterns in Nevada, North Carolina, and Pennsylvania illustrate a series of unusual relationships between candidate support and voter turnout. Using color-coded heat maps, Taylor asserts that his group has discovered statistical distortions similar to those seen in countries with a reputation for fraudulent election practices, such as Russia and Uganda. Using these maps, Taylor alleges that up to 190,000 votes cast in Pennsylvania may have been algorithmically shifted, which would be more than enough to flip the state.
To lend credibility to these claims, the ETA circulated a working paper by the University of Michigan political scientist Walter Mebane that used statistical techniques to examine Pennsylvania’s 2024 election results. Mebane told The Atlantic that while he was aware the group had used his public methodology and data models, he had not reviewed their findings and did not endorse their conclusions.
Hellbound and Down 