An upstart polling firm that had been cited for months by some of the most respected election data analysts in American media was exposed this week as a fabrication. Wedgewood Polls, a company whose surveys were incorporated into polling aggregators run by The New York Times and Nate Silver’s Silver Bulletin, admitted on Tuesday that virtually all of its data had been invented. The confession came only after FiftyPlusOne, a polling and analysis website, pressed the company about apparent inconsistencies in its numbers. In a statement posted to X, Wedgewood called the deception an “experimental project” designed to test how easily a fake polling operation could gain credibility. “We 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 scrutiny exists in media and public commentary?” the statement said. “The short answer: Far too easy.” The admission rippled through the political data world, raising fresh questions about the vulnerability of modern election coverage to manufactured information. For months, Wedgewood’s polls had been shared across political social media and featured in at least some aggregate models covering some of the year’s closest races. The data appeared timely, plausible and, crucially, useful to journalists and analysts hungry for signals in a high-stakes election year. But behind the veneer of methodological respectability, the firm was reportedly inventing numbers and publishing them under the guise of independent research. The episode has left polling professionals and campaign operatives struggling to understand both the motives behind the operation and the broader weaknesses it exposed.
The mystery surrounding Wedgewood’s creators deepened in the hours after the confession. An email sent to an address previously associated with the company bounced back as undeliverable, and Wedgewood’s X account was deleted shortly after the statement was published. The people responsible did not come forward publicly, leaving unanswered questions about who funded the operation, how long it had been active and what they hoped to gain from the so-called experiment. In its statement, Wedgewood claimed to have conducted “one fully legitimate, crowdfunded” poll in Nebraska, but it acknowledged that the vast majority of its other data was fabricated. That concession did little to reassure a data community already shaken by a wave of polling malpractice. Earlier this year, a firm called Medium Strategies acknowledged that it had fabricated survey data in several high-profile races, including the Wisconsin Democratic gubernatorial primary and the Los Angeles mayoral primary. Medium Strategies also described its fake polling as a “social experiment,” and the incident set off a furious debate among political operatives about the ease with which phony numbers could circulate. But the Wedgewood case appears different in important ways. Medium Strategies’ data was largely excluded from the most prominent polling aggregators, limiting its real-world influence. Wedgewood, by contrast, had been woven into the work of major outlets and analysts, suggesting that its methods were sophisticated enough to pass at least initial scrutiny. That distinction has made the episode more alarming to those who rely on polling averages to understand the political landscape.
Exactly how Wedgewood managed to fool so many gatekeepers is now the subject of intense scrutiny. FiftyPlusOne, which first reached out to Wedgewood in June after the company appeared suddenly in the polling ecosystem, said it inquired about the outfit’s background and methodology. Wedgewood replied by identifying its leadership, detailing its start-up capital and explaining its future plans. The response apparently satisfied FiftyPlusOne’s initial concerns, at least for a time. The New York Times likewise said in an editor’s note published this week that it had contacted Wedgewood before including the firm’s data in its polling aggregates. According to the Times, Wedgewood “answered a series of questions regarding the identities of the pollsters; the methods they were using to draw their sample, contact respondents and weight their polls; and the source of their funding.” On its now-deleted X account, Wedgewood took additional steps to cultivate an air of legitimacy, publicizing its email address and soliciting sponsorship money to fund its alleged surveys. That effort convinced even some skeptical data professionals. Zachary Donnini, head of data science at PollHub, said he personally donated money to help fund Wedgewood’s Nebraska poll out of a desire to support young, upstart polling firms. The fact that an industry insider would contribute his own money to a fabricated operation underscores the degree to which Wedgewood’s founders understood the culture of political data: they knew how to speak the language of polling, how to frame their work in methodological terms and how to build trust through transparency rituals. They also understood that in the fast-paced world of campaign news, few journalists have the time or expertise to deeply audit every survey that crosses their desks.
The consequences of Wedgewood’s experiment were not confined to methodological debates or aggregator corrections. Michigan state Senator Mallory McMorrow, who had been running for the Democratic nomination for governor before dropping out before the primary, said the fake polling numbers had a direct and damaging effect on her campaign. After Wedgewood published a poll in late June showing her support collapsing into single digits, McMorrow said her fundraising “dried up” and reporters began calling her about whether she was considering dropping out of the race. “Sure, a point was proven. Polling is broken,” McMorrow wrote on X. “But tens of thousands of people poured their hearts, time, energy, hard-earned money into something they believed in – for a year-and-a-half, and it was treated like a game. Our volunteers gave their all. And then some. I felt like I had let so many of them down.” Her account illustrates how a single piece of fabricated data, even if it does not dramatically shift a polling average, can alter the narrative around a campaign and shape the behavior of donors and the media. A low-quality or fraudulent poll, when amplified by trusted aggregators, can become a self-fulfilling force: it convinces insiders that a candidate is struggling, which in turn makes fundraising harder, which in turn makes the candidate actually struggle. In that way, Wedgewood’s experiment was never harmless. It had real-world victims, including candidates, staff and volunteers who made personal sacrifices based on a political landscape that was, in part, fiction. Major aggregators, including the Times, FiftyPlusOne and the Silver Bulletin, have now removed Wedgewood’s polls from their databases, but the emotional and political damage cannot be undone by a correction.
The Wedgewood scandal lands at a particularly fraught moment for the polling industry, which has spent years defending its credibility against accusations of bias, inaccuracy and malpractice. Recent election cycles have seen spirited public arguments over polling misses, response-rate declines and partisan weighting decisions. The emergence of fake polling firms has added a new and disturbing layer to those debates. In addition to the Medium Strategies episode this summer, a small Florida firm admitted to changing the stated vote choices of certain respondents to better fit the pollster’s “read of the race.” Another firm, SoCal Strategies, which failed to publicly name its pollsters, went dark this week after an inquiry from The Washington Post. The cumulative effect has been corrosive. One Democratic pollster, who spoke on condition of anonymity to discuss the state of the business, called the fake polling episodes a “slap in the face to the entire industry.” The pollster acknowledged that public skepticism of polling is nothing new, but argued that there must be a baseline level of trust between researchers, the media and the public. “These bogus firms are undermining trust and making our jobs harder and your jobs harder,” the pollster said. “Let’s be clear: these are ‘social experiments’ or whatever they’re calling them. It’s misinformation.” The comment reflects a growing fear among legitimate pollsters that the actions of a few bad actors could cause lasting damage to an enterprise already struggling to convince the public of its value. If voters cannot trust the polls they see in the news, they may reject not only bad data but also the carefully conducted surveys that remain important tools for understanding public opinion.
Institutional responses to the Wedgewood revelation have so far been measured. The American Association for Public Opinion Research, a trade organization that helps set standards for survey research professionals, did not respond to a request for comment on Wedgewood Polls. But after the Medium Strategies episode earlier this year, the group’s president, Mary Losch, issued a statement warning that the production and distribution of fabricated data undermines “trust in public opinion and election polling.” Losch emphasized that legitimate polling organizations are transparent and ethical, and that fraudulent activity should not be confused with serious survey research. That message has taken on new urgency now that a second company has been caught manufacturing data. The Wedgewood episode is a cautionary tale about the ease with which misinformation can infiltrate the information ecosystem when verification mechanisms are weak. It also raises uncomfortable questions about the incentives that allow fake polls to flourish. Poll aggregators compete for traffic and influence; journalists crave fresh, credible data; campaigns monitor every number to calibrate their strategy. Each of those actors has a reason to accept a plausible poll without sufficiently probing its origins. Wedgewood’s founders exploited those incentives, and in doing so they demonstrated that the system is only as trustworthy as the checks and balances built into it. Whether those checks will be strengthened remains to be seen. But the damage to public confidence, already fragile after years of polling controversies, may be lasting. For every legitimate pollster working hard to produce reliable information, the rise of fabricated firms represents not just an annoyance but an existential threat to the credibility of the entire field. Wedgewood may be gone, but the questions it raised are only beginning.