The political landscape is grappling with a fresh wave of fabricated survey data after investigators revealed that a newly formed polling company manufactured almost all of its published election numbers. The disclosure follows a similar controversy earlier in the year, highlighting persistent vulnerabilities in how media outlets, data aggregators, and political campaigns vet incoming research.
Wedgewood Polls, a polling entity that emerged during the current election cycle, admitted that only one of the 17 polls it published this year was genuine. The deception came to light after the poll-analysis site FiftyPlusOne investigated the firm’s output. Following the probe, Wedgewood claimed the exercise was designed as a prank or sting to test media scrutiny and demonstrate how easily an entirely fabricated firm could gain traction and be taken seriously.
Before the exposure, Wedgewood’s numbers were widely circulated. Major data platforms and media organizations, including , Decision Desk HQ, Vote Hub, and election analyst Nate Silver, incorporated the firm’s figures into their election tracking. The polls were also cited on television broadcasts and referenced by candidates, such as Mississippi Democrat Scott Colom during his U.S. Senate race against Republican Senator Cindy Hyde-Smith.
While the fake surveys occasionally showed favorable results for both major parties and various factions, they disproportionately favored Democratic candidates and boosted the apparent electability of democratic socialist contenders. Silver’s forecasting model adjusted to become more favorable to Republicans after the fraudulent data was purged from its calculations.
The impact of the bogus numbers was particularly acute in competitive Democratic primaries, where questions of electability often dictate campaign momentum and fundraising. In Michigan, campaigns for primary candidates Haley Stevens and Mallory McMorrow were significantly disrupted. McMorrow experienced a sharp decline in fundraising and support after a June Wedgewood poll placed her in the single digits, ultimately leading her to drop out of the race. Allies noted that donors pulled back funding based on the survey at a time when campaigns were facing heavy spending disparities.
Reflecting on the fallout, McMorrow stated on social media that donations dried up, volunteers sought guidance amid plunging numbers, and reporters frequently pressed her to exit the contest. A SuperPAC fundraiser confirmed that the fabricated survey was utilized by donors to justify withdrawing financial support. Meanwhile, Abdul El-Sayed ultimately won the primary contest, though by a considerably narrower margin than the misleading polling had projected. Liberal columnist Bill Scher had previously raised alarms months earlier regarding pollsters with questionable histories aiding El-Sayed in Michigan and Graham Platner in Maine.
The incident marks the second instance of fabricated polling data surfacing this year. In August, Median Strategies was exposed as bogus after publishing a survey claiming that socialist Francesca Hong held a 23-point lead in the Democratic primary for Wisconsin governor. Hong ultimately lost the primary. Historical precedents include a prominent 2010 scandal involving Research 2000, which published fraudulent surveys for a progressive blog before facing a fraud lawsuit from the site’s founder.
Industry observers note that the growing financial stakes tied to prediction markets have amplified the audience and incentive for polling data. Analysts emphasize that while established polling institutions possess verifiable track records and known methodologies—even when individual surveys prove inaccurate, such as a prominent Iowa poll in late 2024 that preceded a lawsuit later dismissed by the courts—brand-new entities with unverified backgrounds present significant risks for voters, journalists, and campaigns alike.
