Every few days, another poll shows Democrats romping toward a midterm blowout.
Cable hosts read the numbers like scripture. Candidates who trailed in them got told to pack it up.
Now one pollster behind those headlines has confessed to something that should make every voter question the scoreboard.
The Pollster With No Names, No Phone Number, and No Real Data
A shadowy outfit called Wedgewood Polls admitted in October 2026 that it made up the vast majority of the surveys it ever released.
This wasn’t a rounding error or a bad sample. The firm made the numbers up.
The firm posted a statement claiming the whole thing was a social experiment to test how easily a phony pollster could fool the political press. Its answer to its own question was blunt: “The short answer: Far too easy.”
Wedgewood bragged that its numbers landed in the polling databases at The New York Times, FiftyPlusOne, and VoteHub. “Our fabricated numbers were given instant credibility across national media platforms,” the statement read, and it named MS NOW as one outlet that cited the figures on air and in its reporting.
According to Wedgewood, only one poll was real. That was a Nebraska survey the firm said it used as a baseline to compare against what it called synthetic data.
But the timing of the confession was awfully convenient. Cooper Burton, a data scientist at FiftyPlusOne, said his outfit had contacted Wedgewood for comment that same morning ahead of a planned investigation. “This is just an attempt to save face after they realized we caught them,” Burton wrote. FiftyPlusOne pushed its report out early and disputed Wedgewood’s claim that nobody vetted the firm. It said Wedgewood got in the door because experienced pollsters ran it.
So somebody did vet it. And the fake numbers still sailed straight into the national bloodstream.
The firm’s setup raises even more eyebrows. Straight Arrow News reported that Wedgewood’s website listed no employees, no phone number, and no location, and lived on a free Google Sites page. The firm said online harassers had doxxed one of its two members, so both stayed anonymous. The New York Times polling page did note that Wedgewood failed to meet its criteria for selected pollsters. And it listed the numbers anyway.
How One Phony Survey Knocked Mallory McMorrow Off Her Feet
Michigan’s Democrat Senate primary turned into a three-way slugfest for the right to face Republican Mike Rogers for an open seat.
In late June 2026, Wedgewood released a Michigan survey showing Abdul El-Sayed, the progressive backed by US Senator Bernie Sanders (I-VT), at 46 percent. Michigan State Senator Mallory McMorrow sank into single digits in the same poll.
McMorrow has since described what happened next. She said her internal data told a completely different story. But that didn’t matter. Donations dropped sharply. Volunteers who had poured 18 months into the campaign lost heart. And reporters started asking when she planned to drop out.
McMorrow said an earlier Mitchell poll with its own problems had already helped build the narrative, and the fake survey gave it fresh momentum. El-Sayed started to look inevitable. McMorrow eventually left the race, and she now backs El-Sayed as the Democrat nominee against Rogers.
That means a firm that admits it invented its numbers helped shape who Michigan Democrats nominated for one of the most closely watched Senate seats in America.
Let that sink in for a second.
And Wedgewood didn’t stop at Michigan. Urban Milwaukee reported that the firm released a Wisconsin poll showing far-left gubernatorial candidate Francesca Hong at 30 percent when legitimate surveys had her well below that. Its social media account billed itself as “Progressive but unbiased” and cheered on candidates from the hard-left wing of the party.
Who actually ran it? Nobody knows. One Milwaukee writer floated the theory that it was a Republican front posing as a liberal shop. Others see an operation built to boost the socialist wing of the Democrat Party. The fact that nobody can say for sure is the whole problem.
Every Rosy Democrat Poll Just Earned a Second Look
Wedgewood isn’t the only bad apple. Earlier this year, an outfit called Median Strategies put out polls on the Los Angeles mayoral race and Wisconsin’s Democrat governor primary, then announced that its entire operation had been a social experiment too. Local TV newscasts and online outlets had already run with its numbers.
Then FiftyPlusOne permanently removed a firm called The Public Sentiment Institute from its site. The company had admitted shifting some respondents’ answers toward Florida Republican gubernatorial candidate James Fishback, whose campaign had paid it about $1,500. Rot like that doesn’t care about party labels, and voters on the Right should want every one of these shops dragged into the daylight.
But look at where the bad numbers in Michigan and Wisconsin pointed. The fake surveys that got the most traction there propped up the hardest-left candidates in Democrat primaries and fed the storyline that the party has momentum heading into November.
That raises the obvious question. How many of the polls showing Democrats with eye-popping leads over Republicans rest on numbers just as hollow?
Nobody can answer that yet, and it would go too far to say every rosy Democrat poll is fake. Most come from real firms doing real interviews. But a polling average is only as good as the numbers pollsters feed into it. The same aggregators that swallowed Wedgewood feed the averages networks flash on screen every night. FiftyPlusOne now serves as the exclusive national partner for MS NOW’s midterm polling graphics, which is the same network Wedgewood boasted about fooling.
President Trump’s own pollster flagged trouble in the Michigan numbers well before Wedgewood came clean. Tony Fabrizio, who polls for Rogers, circulated a memo in August 2026 arguing that public surveys showing El-Sayed ahead oversampled Democrat-leaning voters, Politico reported. His complaint concerned sampling at real firms rather than fabrication, but it pointed at the same weak spot.
Polls don’t just measure an election. They move money, steer volunteers, and tell reporters which candidates to take seriously, as McMorrow learned the hard way. A fake number a firm drops in June can bury a campaign long before anyone casts a single vote.
McMorrow now urges candidates to knock on doors instead of obsessing over the toplines. Coming from someone a phony survey helped push out of the race, that’s about as close to common sense as Democrats get these days.
Republicans in tight races this fall should take the same advice. A poll showing a GOP candidate down eight points might reflect reality. It might also come from a Google Sites page run by people who won’t give their names.
And voters deserve to know which is which before November, not after.
Sources: Raw Story; Straight Arrow News; Urban Milwaukee; The National News Desk (WJLA); Decision Desk HQ (Wedgewood Polls Michigan survey, June 2026); NewsNation; FiftyPlusOne; The Daily Wire; Political Wire/Politico