2026 Midterm Polling: Missing Voters and Fake Data Cloud Outlook

Polling experts face a difficult task this election cycle: measuring a electorate that increasingly avoids them and a media environment rife with fabricated data. Charles Franklin, director of the Marquette Law School Poll, told The Monitor that these structural issues have created uncertainty in the lead-up to the November 3 midterms.

Franklin traced the current challenges back to the 2016 election. He noted that Donald Trump mobilized a segment of the electorate estimated at 4% to 6% who historically did not vote or answer surveys. These voters, Franklin explained, were often skeptical of government and media, leading them to ignore pollsters. This dynamic introduced a systematic bias where Trump supporters were underrepresented, causing polls in Wisconsin, Michigan, and Pennsylvania to incorrectly show him trailing.

Since then, the industry has shifted methods. Marquette now conducts most polls online, sending emails and texts to a sample drawn from state registered voter lists. Franklin said that if four attempts to contact a participant fail, staff still place a telephone call. This hybrid approach results in roughly 5% to 10% of the sample being reached by phone, a drop from the 100% telephone contact rate used prior to 2022.

This cycle introduced a new complication: fake polls. Two organizations, Median Strategies and Wedgewood Polls, released data that was widely cited before being exposed as fabricated. Franklin said Median Strategies gained attention when the Los Angeles mayor retweeted their numbers, while Wedgewood was included in The New York Times average. Both organizations later admitted to the deception, describing it as a social experiment to test media vetting.

The impact of these fakes was tangible. In Michigan, Democratic state Sen. Mallory McMorrow dropped out of the U.S. Senate primary after a fake Wedgewood poll showed her in single digits. She cited a drying up of fundraising in a statement. Franklin observed that such incidents highlight the vulnerability of campaigns and prediction markets like Kalshi and Polymarket to unverified data.

Franklin also addressed the Wisconsin Democratic primary, where polls showed democratic socialist Francesca Hong leading by a wide margin. She finished with 39.3% of the vote, narrowly losing to moderate Democrat David Crowley. Franklin attributed the shift to a chaotic final month involving candidate dropouts and endorsements. He noted that his polls identified undecided older voters who were less favorable toward democratic socialism, though the data did not confirm they would support Crowley.

With the November 3 vote approaching, Franklin said pollsters are working to refine their methods for 2028. He emphasized that the current gap in data means experts must remain cautious as they interpret results from the real polls currently suggesting a competitive national race.

The Monitor