Two months before Wedgewood Polls entered her life, Mallory McMorrow was tied for first in Michigan’s Democrat Senate primary at 24%, according to a legitimate Emerson College poll. Then, on June 24, Wedgewood released its own survey. Abdul El-Sayed led with 46%, Haley Stevens had 39%, and McMorrow was down to 9%; for good measure, the release cast her as the only Democrat who couldn’t beat Mike Rogers in November. She left the race eleven days later.
None of it was real. Wedgewood now admits that nearly everything it published this cycle was fabricated, and the New York Times has quietly pulled 15 Wedgewood surveys from its polling tracker. McMorrow says the money dried up, her volunteers lost heart, and reporters started asking when she would withdraw. The woman who ran the super PAC backing her says donors pointed to the Wedgewood numbers when they closed their checkbooks.
That audience is where a fake poll does its work. It never has to change a single vote at the ballot box because the people who decide whether a candidate is viable read the polls long before voters do: donors, endorsers, party committees, reporters, and finally the candidate herself. Four firms exploited that fact in 2026. Four firms exploited that fact in 2026, and the Department of Justice and the state attorneys general, beginning with Texas Attorney General Ken Paxton, should treat them as the fraudsters they are.
Wedgewood was caught when Nate Silver’s FiftyPlusOne confronted it with crosstabs that couldn’t exist and got an admission that most of its releases were made up, which the firm later repackaged as a “social experiment.” Wedgewood named Jack Sims and Andrew Prater as its operators. Sims, it turns out, has real polling experience, and that detail should end any talk of a prank. Somebody who knew exactly what a credible poll looks like built a product designed to pass as one.
SoCal Strategies, a Florida business connected to the same Jack Sims, listed Wilson Rochford as its lead pollster. Rochford’s actual role is disputed, and investigators have been told someone may have been using his identity. A prosecutor with subpoena power ought to settle that question, since a blogger with a browser cannot. SoCal vanished from the internet within hours of Legal Insurrection and others asking about its ties to Wedgewood, and FiftyPlusOne stopped collecting its polls on October 22.
The Public Sentiment Institute (TPSI), run by head pollster Lester Tellez alongside Dustin Smith, was banned after FiftyPlusOne found an undisclosed $1,575 payment from James Fishback’s Florida gubernatorial campaign, along with an internal postmortem describing how Jay Collins supporters were reassigned to Fishback. Rahil Prakash, a 21-year-old recent graduate, created Median Strategies and told reporters he built the whole operation alone with AI. He invented polls in Los Angeles, Wisconsin, and Nevada. Then he watched Karen Bass’s mayoral campaign promote his fiction as fact, which tells you how little curiosity a Democrat campaign brings to a number it likes.
The fakes hit the Democrat primaries first, and they negatively impacted moderate Democrats. Wedgewood’s July 21 Wisconsin survey showed the progressive Francesca Hong at 38%, Mandela Barnes at 30%, and David Crowley at 9%, and Hong’s campaign put the numbers in a press release the next day. On the prediction market, Crowley’s price fell from roughly 23% to 13% within 15 minutes of the release. David Crowley even suspended his campaign and endorsed his opponent. He reentered the race only because she dropped out. In the end, he won. A weaker candidate would not have survived the fortnight.
Mississippi got the same treatment. Wedgewood claimed Cindy Hyde-Smith led Scott Colom by just 52% to 48%, Colom’s campaign trumpeted it, and the Democrats and their willing accomplices in the drive-by media spent a week running breakthrough stories about a red state in play. Silver’s own analysis found that Wedgewood’s numbers ran more Democratic than his averages and were especially generous to the furthest-left (DSA & Islamic Marxist) candidates in several contests. The lies leaned left.
Later the operation turned its guns on Republicans. TPSI’s Florida numbers nudged Collins voters toward Fishback for a paying client nobody disclosed. SoCal released a Texas Senate runoff survey of 700 likely voters showing Ken Paxton at 57% and John Cornyn at 35%, and Capitol Inside promptly reported that Paxton was approaching a decisive victory. SoCal’s Kentucky survey of 450 likely voters put Ed Gallrein at 49% and Thomas Massie at 42%. Jim McLaughlin, the president’s pollster, calls these suppression polls. Tom Bevan laid out in RealClearPolitics why RCP refused to carry them, even as the Times and FiftyPlusOne did.
The forgers walked into an already poisoned environment. CNN has conceded on air that Republicans outperform the so-called legitimate polls by roughly 5 points, and Republican voters stopped believing them. Here in Texas, pollsters admit openly that they cannot reach Republican voters because Republican voters hang up on them. Then there are a dozen or so polling shops, financed and staffed by partisan organizations, that still describe themselves as independent. In an industry like that, the honest practitioner is hard to tell from the operative, and the four fake firms sit at the far end of a spectrum the profession tolerated for years.
A voter who already suspects the polls are rigged against him will not be surprised by a bogus 22-point deficit. It will discourage him, and a discouraged voter stays home. In a primary, a fake poll ends candidacies, while in a general election it drains turnout from exactly the voters who had learned to distrust the polls in the first place.
Silver insists the damage was small, and he has a number to back him up. Pulling Wedgewood out of his Michigan average moved El-Sayed’s lead by 0.6 points, from 3.6 to 3.0. The figure is accurate and also beside the point, since the average was never what the forgers were aiming at. The donor was. So was the reporter who asked McMorrow when she planned to drop out.
Deleting a fraudulent poll from an average cannot undo the campaign decisions and public perceptions it helped create. McMorrow cannot un-withdraw. Crowley’s donors cannot get back the week they spent hedging, and no correction on a polling tracker returns the money a Mississippi Democrat raised on a lie. Real money changed hands, real decisions were made, real markets moved, and harm like that is exactly what criminal law exists to address.
Attorney General Todd Blanche should begin with the wire fraud statute, 18 USC §1343. It carries up to 20 years in prison and a $250,000 fine per count, and every email, every crowdfunding solicitation, and every invoice sent across state lines is its own count. Wedgewood asked the public for money to pay for polling it never conducted. In Kousisis, decided in 2025, the US Supreme Court settled that fraudulent inducement, which means lying to get someone to hand over money, is wire fraud even when the victim cannot show a net loss. Conspiracy under §1349 carries the same 20 years for anyone who knowingly joined the scheme.
Where an account connected to the operation traded before a release, as the trade analysts at BETINT allege happened with SoCal, prosecutors should add 7 USC §13(a)(2), which makes it a crime to manipulate a regulated contract market and carries up to 10 years and a $1 million fine. The CFTC’s February 2026 advisory on event-contract fraud gives civil enforcers a running start. TPSI’s secret work for a campaign also implicates the Federal Election Campaign Act. When a campaign pays for a poll that is then released as independent research, the campaign has made an in-kind contribution it never disclosed, and under 52 USC §30109(d), a knowing and willful violation involving more than $25,000 in a year carries up to 5 years.
Texas does not have to wait on Washington, and General Paxton should move now. Senate Bill 12, effective December 4, 2025, restored his authority to prosecute election crimes, though the quicker route runs through the Deceptive Trade Practices Act. That law forbids falsely representing the sponsorship, characteristics, or quality of a service, and §17.47 authorizes injunctions, civil penalties of up to $10,000 per violation, and restoration of every dollar obtained by deception. Sections 17.60 and 17.61 give the attorney general civil investigative demands, which work like subpoenas issued before any lawsuit is filed. He could compel the bank records, vendor contracts, and survey identifiers tomorrow.
On the criminal side, Penal Code §32.32 punishes a false written statement made to obtain property and grades it like theft, so a scheme that took in between $2,500 and $30,000 is a state jail felony carrying 180 days to 2 years, and one that took in more than $30,000 is a third-degree felony carrying 2 to 10 years. Forgery under §32.21 reaches invented respondent records passed off as genuine. Deceptive business practices under §32.42 adds a Class A misdemeanor and up to a year in county jail.
Paxton stands in a peculiar spot, because his own runoff was the subject of SoCal’s 57% to 35% fiction. The fastest way to prove the attorney general’s office is not an instrument of his campaign is to prosecute the people who tried to help it, and Government Code §402.028 lets him bring in the district attorneys so the operators’ home counties can file the charges.
Somebody will object that lying is protected speech, and the US Supreme Court did say as much in Alvarez. The Court has never said that taking money for a service you never performed is protected, though, and it has never said so about manipulating a regulated market or about funneling secret campaign contributions through a fake research shop either; that is why every charge above rests on the money and the market instead of the lie itself. The First Amendment defense will definately collapse the moment a prosecutor produces the first invoice for interviews nobody conducted. Juries understand the distinction, even when pollsters pretend they don’t.
A poll is public infrastructure, like a bridge inspection or a bank audit, and a republic that lets forgers counterfeit it will eventually get the candidates the forgers prefer. We prosecute the man who prints fake currency even when the economy absorbs his bills. The next printer will be better at it, and the one after him will be a foreign goverment.
The 2026 fakes were caught because a few statisticians noticed crosstabs that could not exist. Nobody should count on the next batch being that careless.
If you enjoy my work, subscribe x.com/amuse/creator-subscriptions/subscribe
Alexander Muse is a Fellow at the John Milton Freedom Foundation and publishes political analysis at amuseonx.com. Readers who identify errors are invited to contact me directly. Each op-ed is edited for grammar and clarity using AI in partnership with Grammarly. Data provided in a sponsored partnership with Polymarket.
Brilliant legal assessment. AG Paxton, are you reading this?
Finally an easy to understand way to understand the intentionally messy, tangled web of massive fraud in our elections. White collar crime….but I believe if these max sentences were increased to mandatory 10 yrs for each illegal act you could almost shut down this crime in its entirety. No? Let’s just try it for 10 years because as it is, the hand slapping, gentlemanly way of doing business is neither an impressive deterrent nor a punishment equal to the crime of assaulting voters and violating the sacred privileged rights of our trust in voting.