A polling organization known as Wedgewood Polls acknowledged on October 17 that all but one of its surveys were fabricated, describing the effort as an experiment to expose vulnerabilities in the polling industry. The company stated its surveys were designed to demonstrate how easily unverified data could gain credibility in media and public commentary. Decision Desk HQ and VoteHub, two major poll aggregators, removed 17 Wedgewood Polls from their datasets following the admission, though neither reported changes to their projected seat totals in the U.S. Senate or House.
The revelation has raised concerns about the potential impact of false polling data on campaign fundraising, donor confidence, and candidate perceptions in competitive races. Wedgewood Polls conducted a single legitimate survey in Nebraska to serve as a baseline for comparison, while all other data were generated without real respondents. The fabricated polls were cited by outlets including The New York Times and MSNBC, with several surveys showing Democratic candidates leading in races such as the Michigan Senate primary.
Poll aggregators and media organizations have begun reviewing their vetting processes in response. FiftyPlusOne, an election analysis website, uncovered irregularities in Wedgewood Polls' data, prompting the firm’s admission. The company stated it did not intend to mislead the public or influence elections, emphasizing its goal was to highlight systemic issues in polling oversight and media scrutiny. While the removal of Wedgewood Polls’ data did not alter projected outcomes in major races, the incident has renewed scrutiny of polling accuracy ahead of the midterm elections.
The polling industry has faced criticism for years, particularly since Donald Trump’s 2016 presidential victory, when polls failed to predict his win. This latest controversy underscores ongoing concerns about transparency and accountability in political polling, with experts calling for stricter verification standards to prevent the spread of fabricated data.