Concerns over the integrity of election forecasting have recently shifted focus amid controversies involving online prediction markets and polling practices in the United States. While prediction markets, which allow participants to bet on election outcomes, have drawn calls for increased regulation or outright bans from state regulators, advocacy groups, and some lawmakers, recent developments suggest that polling practices may pose a greater risk to election integrity.

Prediction markets have a long history in American politics, dating back to the early 19th century, with notable venues operating near the New York Stock Exchange. Despite their longstanding presence and general reputation for accuracy, critics have argued these markets could be manipulated to influence voter behavior or provide financial incentives to affect election outcomes. However, academic research spanning 1900 to 2024 indicates these markets remain resilient to political bias, even though the demographics of their participants tend to be skewed. Experts also note that those seeking to manipulate elections for financial gain can often achieve greater returns by trading traditional securities rather than in prediction markets.

In contrast, the polling industry has recently faced significant scrutiny following revelations of fabricated polling data. In the summer of 2026, the firm Median Strategies admitted to producing false polls, notably exaggerating the support of a far-left candidate in Wisconsin's Democratic gubernatorial primary. The company described this activity as a “short-term social experiment” intended to observe the spread of purported polling information in the political ecosystem without independent verification.

More recently, Wedgewood Polls acknowledged that much of its polling was also part of an experiment. This deception appears to have had tangible effects on election outcomes. In Michigan’s Democratic U.S. Senate primary, state Senator Mallory McMorrow withdrew her candidacy after Wedgewood’s fabricated numbers showed her trailing, leading to reduced campaign donations and diminished morale. McMorrow attributed her withdrawal directly to the impact of the false polls. Abdul El-Sayed subsequently secured the nomination and is engaged in a close general election contest for the seat.

The influence of these phony polls extended beyond candidates. Prominent outlets such as The New York Times incorporated Wedgewood’s data into their polling aggregates, and noted analysts like Nate Silver also factored these results into their evaluations. Notably, the fabricated polls tended to depict a narrower Republican advantage in races such as the Mississippi Senate election than prediction markets indicate, which currently give Republicans a roughly 90% chance of victory. This discrepancy has contributed to shifting narratives surrounding the midterm elections, affected donor behavior, and potentially influenced voter decisions.

Instances of fraudulent polling are not unprecedented; in 2006, two individuals pleaded guilty to federal fraud after deceiving campaigns with falsified internal polling data. The recent cases underscore a broader vulnerability within the polling industry. Unlike prediction markets, where no direct evidence links manipulation to altered election outcomes, fabricated polls carry clear proof of undermining electoral integrity. The ease with which malicious actors can disseminate fake polling data raises concerns about the potential for further distortions in future elections.

Experts argue that the polling industry has so far failed to self-regulate effectively and recommend legislative action to safeguard election forecasting. Proposed measures include mandatory auditing of poll results, disclosure requirements regarding pollsters’ affiliations, and financial penalties for deceptive practices. Advocates call for shifting regulatory scrutiny from prediction markets to polling firms, emphasizing that protecting public trust in elections requires addressing the area with demonstrated risks to electoral fairness.