Fabricated Polls Raise Concerns Over Election Integrity

The integrity of political polling has been called into question following revelations that at least two new polling firms, Median Strategies and Wedgewood Polls, have fabricated data. These instances highlight a broader concern about the scrutiny applied to survey results and the ease with which fabricated data can gain traction in the media and public discourse.

Wedgewood Polls, which emerged this election cycle, admitted that only one of the 17 polls it published was based on real data. The firm claimed its intention was to test how easily a fabricated polling operation could be taken seriously. However, its polls were cited by various news outlets and political figures, including The New York Times, Decision Desk HQ, Vote Hub, and Nate Silver's analysis. Mississippi Democrat Scott Colom cited a Wedgewood poll showing him tied in his U.S. Senate race, and the poll was included in FiftyPlusOne's election data. Following an investigation by FiftyPlusOne, which exposed the fraud, Wedgewood admitted its operations were largely a "prank or a sting."

These fabricated polls have often shown favorable results for Democratic candidates, particularly Democratic Socialists, and have been used to bolster their perceived electability. After Nate Silver removed the fake polls from his calculations, his model became more favorable to Republicans. Liberal columnist Bill Scher had previously raised concerns about pollsters with questionable histories, and candidates who lost to figures like Abdul El-Sayed in Michigan have expressed frustration.

Mallory McMorrow, a Democratic candidate who lost to El-Sayed, stated that a Wedgewood poll showing her in single digits significantly impacted her fundraising, leading her to drop out of the race. A SuperPAC fundraiser for McMorrow indicated that donors used the poll to withdraw funding, exacerbating a spending deficit. Neera Tanden, a supporter of Haley Stevens, another Democratic candidate who lost to El-Sayed, noted that some polls favoring El-Sayed over Stevens were either made up or highly partisan.

Gary Peters, the outgoing Democratic incumbent, had voiced concerns that misleading polls could distort primary outcomes. Ultimately, El-Sayed won the primary by a narrower margin than projected by some polls. Historically, fabricated polls have influenced primary contests. In 2010, Research 2000 published bogus surveys for the progressive blog Daily Kos, leading to a lawsuit for fraud. Such polls can shape perceptions of electability, a critical factor in primaries where candidates need to convince voters they can win in the general election.

The issue of fake polls is amplified by the increasing reliance on poll data, with prediction markets now involving significant financial stakes. While skepticism is warranted, the article suggests that legitimate polls, when interpreted cautiously, can still offer valuable insights into electoral possibilities. However, it stresses the importance of vetting pollsters.

Past instances of inaccurate polling have also drawn attention. J. Ann Selzer of the Des Moines Register faced criticism after a late 2024 poll predicted a Kamala Harris victory in Iowa, despite Donald Trump winning the state by a substantial margin. Trump subsequently sued Selzer for fraud, though the lawsuit was dismissed. The article distinguishes between genuinely inaccurate polls and deliberately fabricated ones, the latter of which constitutes fraud.

Experts advise voters and reporters to prioritize pollsters with established track records and known institutions behind them, such as Gallup or university-affiliated polls. In contrast, new polling firms with no verifiable history or institutional backing warrant extreme caution. The advice for the public and media is clear: demand transparency and a history of reliability before trusting pollster data.