Fake polls fooled the nation’s biggest newsrooms for four months

By Wayne Laugesen | Commentary, Wayne’s Word

You can’t make this stuff up. Or can you?

A “polling firm” called Wedgewood Polls published made-up numbers on major races for four months. Newsweek, The Hill, Decision Desk HQ, Nate Silver’s Silver Bulletin, The New York Times’ polling database, and other media passed them along as real.¹

The firm surfaced in June with a real website but no staff names, phone numbers, or addresses. On Oct. 20, researchers at FiftyPlusOne confronted Wedgewood operators with proof their raw data was fake. The fraudsters conceded, saying it was “far too easy.”¹ As if elections are a joke, they called the whole thing an experiment to expose the polling industry’s weak standards. 

Sure. Just as a guy caught with stolen goods might say he was testing a store’s security.

Fun & games till people get hurt

In June, Michigan Democrat Mallory McMorrow was running for the U.S. Senate when a Wedgewood poll put her at 9%. Donations dried up, volunteers were devastated, and reporters urged her to quit. She dropped out on July 5. “I felt like I had let so many of them down,” she said this week.²

In Louisiana, an early Wedgewood poll had John Fleming leading Julia Letlow in the GOP Senate primary. Turns out Letlow won by nearly 14 points.¹

For what it’s worth, the Times had flagged Wedgewood as failing its standards but listed the fake polling firm’s numbers anyway. (The Times has never been a stickler for the truth.) Decision Desk HQ later pulled 17 Wedgewood polls from its averages. 

Ultimately, nobody checked the basics until outside researchers did.

Polls don’t just measure races, they move them

It’s called the bandwagon effect. 

Researchers have documented that when polls show a candidate gaining ground they attract more voters, who jump on the bandwagon.³ Contrarily, supporters of a candidate losing in the polls tend to lose heart and get off the bandwagon. 

Money follows the same signal. Donors treat strong numbers as proof a candidate can win, and campaigns flash favorable surveys to garner big checks. Weak numbers starve a campaign of cash, staff, and momentum. In crowded primaries, party insiders use polls to decide who stays and who goes. One well-timed survey can reshape the field — the poll’s authenticity apparently doesn’t matter.

As a result, polling is anything but neutral. When a number can determine donations, volunteers, headlines, and candidates’ decisions, the temptation to produce the “right” number grows. Wedgewood showed how easily anyone can act on that temptation.

When the polls got it wrong

Fake polls are new. Wrong polls are not.

• 1948, National: Gallup’s final poll had Thomas Dewey beating Harry Truman, 50% to 44%. Truman won, 50% to 45%. The Chicago Tribune’s famous “Dewey Defeats Truman” headline became a monument to polling arrogance.⁴
• 2020, Wisconsin: A week before the election an ABC News-Washington Post poll put Joe Biden up 17 points.⁵ He won Wisconsin by less than 1 point.
• 2020, Maine: Every poll in the RealClearPolitics average showed Sen. Susan Collins losing, one by 12 points. She won by nearly 10.⁶
• 2024, Iowa: Ann Selzer, Iowa’s most respected pollster, released a final poll showing Kamala Harris ahead by 3. Donald Trump won Iowa by 13, a 16-point miss.⁷

Polls are getting worse

Most Americans no longer have landlines, and few answer calls from unknown numbers. By 2018, typical response rates for telephone surveys had fallen to 6%.⁸ Cheap online panels fill the gap, with unrepresentative samples and a growing army of fake respondents.

Meanwhile, trust in the media that hypes these polls has cratered. Gallup found just 28% of Americans trusted the mass media in 2025, a record low, before a slight rebound to 33% this year. Another third say they have no confidence at all.⁹ Considering how many polls miss we should suspect those numbers flatter the press.

Artificial intelligence will make fake polling easier and more convincing. A machine can now generate realistic crosstabs, respondent-level data, and professional memos in minutes. 

The warning

Treat horse-race polls with deep skepticism. They are harder than ever to get right, even by pollsters with pure intentions. They too often pass along information without scrutiny and change the very outcomes they claim to measure. When a number can move money, media attention, and candidates’ decisions, the people who produce and promote it have every reason to shape the story.

Wedgewood wasn’t a curiosity. It was a warning: Let elections, not polls, determine who governs us.

READ THE FULL COMMENTARY AT WAYNE’S WORD

Editor’s note: Opinions expressed in commentary pieces are those of the author and do not necessarily reflect the opinions of the management of the Rocky Mountain Voice, but even so we support the constitutional right of the author to express those opinions.