A US-based polling firm, Wedgwood Polls, has admitted to fabricating much of its data in an incident that adds to growing concerns about the reliability of public opinion surveys ahead of the November midterm elections. The admission came after an investigation by polling watchdog FiftyPlusOne, which questioned the legitimacy of Wedgwood’s surveys used by political campaigns and mainstream media outlets.
Wedgwood Polls stated on social media that its fabricated polls were part of an experiment to explore how easily a fully fabricated polling entity could gain traction and be taken seriously by poll aggregators and media organizations. The group’s data included polls on various midterm races, affecting perceptions of candidates such as Michigan Senate contender Mallory McMorrow, whose reported low support in a false Wedgwood poll reportedly led to a drop in donations and volunteer activity.
This is the second major polling scandal in recent months following the admission of another group, Median Strategies, which also disclosed that it produced false polling data as a similar experiment. While most poll trackers avoided Median Strategies, Wedgwood managed to have more than a dozen of its polls included by widely followed aggregators such as those maintained by the New York Times and FiftyPlusOne.
FiftyPlusOne’s inquiry was triggered by a suspicious poll from Wedgwood showing a narrower-than-expected lead for Republican Senator Cindy Hyde-Smith in Mississippi shortly after a prominent analyst changed the race rating. When FiftyPlusOne received the raw data for inspection, several irregularities emerged. The data appeared unusually clean and closely matched population demographics without implemented quotas, raising doubts. Further checks revealed discrepancies between Wedgwood’s data and an independently conducted poll using the same questionnaire and sample provider, Pollfish.
Metadata analysis showed inconsistencies in respondent identifiers and timestamps in Wedgwood’s data, which did not correspond to typical survey response behavior. These anomalies led FiftyPlusOne to press Wedgwood for explanations, culminating in the firm’s public confession.
The incident has raised questions about how poll aggregators verify the legitimacy of new polling firms. Some, like RealClearPolitics, declined to incorporate Wedgwood polls, citing lack of track record, missing staff details, and incomplete online presence. FiftyPlusOne has pledged to enhance vetting practices, including personal interviews and deeper reviews of microdata, to prevent similar fraud in the future.
Experts note that the advancement of AI tools has made it easier for bad actors to generate plausible yet fraudulent polling data and documentation. This has complicated the process of distinguishing genuine surveys from fabrications.
The cases of Wedgwood and Median Strategies highlight a recurring theme in 2026 polling fraud: anonymity. Both firms provided limited or no public information about their staff, evading accountability. In contrast, another firm, Quantus Insights, recently increased transparency by publishing names and photos of its pollsters. However, SoCal Strategies, a polling group linked to Wedgwood personnel, remains the only major firm publicly unidentified, and ceased its online presence following the scandal.
The developments underscore ongoing challenges within the US polling industry amid an already polarized political landscape, prompting calls for improved transparency and stricter vetting to maintain public trust ahead of critical elections.