Election analysts have uncovered a second major polling fraud this year after Wedgewood Polls, a firm that emerged during the current election cycle, admitted that only one of the 17 surveys it published in 2026 reflected real data.
The fabrication was identified by the poll-analysis site FiftyPlusOne, which had been including Wedgewood’s numbers in its election tracking. After being caught, Wedgewood claimed its fake surveys were actually a deliberate test, saying it wanted to “test a simple hypothesis: How easy is it for an entirely fabricated polling firm to gain traction, be taken seriously, and expose how little scrutinization exists in media and public commentary?”
Before the exposure, Wedgewood’s polls had already been cited by The New York Times, Decision Desk HQ, Vote Hub, statistician Nate Silver, and the network MS NOW. Mississippi Democrat Scott Colom touted a Wedgewood poll showing him tied with Republican Sen. Cindy Hyde-Smith in the state’s US Senate race. Silver’s election model shifted somewhat more favorable to Republicans after he removed the fake Wedgewood data from his calculations.
This follows an earlier case in August, when Median Strategies, another newly formed pollster, was exposed for claiming socialist candidate Francesca Hong held a 23-point lead in Wisconsin’s Democratic gubernatorial primary. Hong went on to lose that race.
Impact on Michigan’s Senate primary
The fallout has been most visible in Michigan, where Democrats Haley Stevens and Mallory McMorrow both lost a Senate primary to Abdul El-Sayed. Liberal columnist Bill Scher had warned months earlier about pollsters with questionable histories boosting El-Sayed as well as Graham Platner in Maine.
McMorrow dropped out of the race after her poll numbers collapsed, including a June Wedgewood poll that showed her in single digits. She wrote on X that “donations dried up, volunteers were devastated and pleading for guidance on what to do,” and that “reporters hounded me with calls to drop out.” A fundraiser for a McMorrow-aligned super PAC said the poll “was used by several donors to pull back funding at a time we were being outspent by $80 million.”
Neera Tanden, who had supported Stevens, wrote: “Man the pollsters that showed Haley behind over the summer were either made up or highly partisan. And a lot of smart people fell for the op.” Outgoing Democratic Sen. Gary Peters had voiced concern during the primary that misleading polling was distorting the race. El-Sayed ultimately won by a much narrower margin than the earlier polls had projected.
A familiar pattern
The episode recalls a 2010 scandal involving Research 2000, a firm that published fabricated surveys for the progressive blog Daily Kos. Daily Kos founder Markos Moulitsas later sued the company for fraud. In both cases, fake polling was used to portray progressive primary candidates as more electable than they turned out to be — a key argument in primaries where opponents argue a candidate can’t win in the general election.
The use of prediction markets, where real money is wagered on poll-driven outcomes, has raised the stakes for fake polling to reach wider audiences.
Not every polling failure involves fraud. Ann Selzer, the longtime Iowa pollster for the Des Moines Register, retired after a late-2024 poll wrongly predicted a Kamala Harris win in Iowa; Donald Trump won the state by 13 points and later sued Selzer, a suit that was dismissed. Analysts distinguish that kind of polling miss from Wedgewood’s conduct, in which most of its published surveys were never conducted at all.
