A polling organization that published attention-grabbing numbers in political races in Wisconsin and California has admitted there was one major problem with its surveys: They were never real.
Median Strategies told the Los Angeles Times that the organization was actually a “social experiment” designed to see how easily supposed polling data could spread through political circles and the media without anyone independently verifying it.
The admission came after Median published a spectacularly inaccurate poll of Wisconsin’s Democratic gubernatorial primary. The organization had far-left candidate Francesca Hong leading by 23 percentage points.
Hong subsequently lost the primary.
Median initially responded to the miss as though it had simply suffered the kind of polling failure that occasionally happens to legitimate firms.
“We got Wisconsin wrong- and by quite a margin too,” the organization said.
But there was nothing to get wrong because, according to Median’s subsequent admission, there had been no genuine polling behind the numbers in the first place.
“Median Strategies was created as a short term social experiment to examine how purported polling information could enter and spread through the political information ecosystem without independent verification,” the organization told the Los Angeles Times in an email.
“We are not seeking publicity or attribution for the individuals involved and are declining interviews,” it added.
That explanation represented quite a reversal from the image Median had cultivated publicly.
“Polling has a credibility problem,” the organization previously declared on social media. “Ours is the fix: no partisan clients, no campaign or party money, numbers that go wherever the data takes them even when it’s inconvenient. Trustworthy polling, rebuilt.”
Those claims of rebuilding trust were themselves part of an operation publishing invented numbers.
And Wisconsin was not the only race involved.
After Hong’s defeat, Median released another purported poll concerning the Los Angeles mayoral race. This one showed incumbent Mayor Karen Bass leading City Council member Nithya Raman by nearly 12 percentage points.
That poll was fictional as well.
So this is pretty wild: A polling firm that came out of nowhere to show Francesca Hong with a 23-point lead over David Crowley in the Wisconsin primary has shut down and said its whole operation was bogus pic.twitter.com/QTmlXkvGPC
— Brett LoGiurato (@BrettLoGiurato) August 18, 2026
Median also published polling concerning Nevada’s gubernatorial race before ending the project.
The episode demonstrates a basic vulnerability in modern political coverage: A professional-looking organization can publish precise percentages, wrap them in the language and presentation of conventional polling, and potentially watch the numbers travel before anyone establishes whether respondents were ever surveyed.
Chris Wilson, a Republican pollster who now heads public opinion firm EyesOver, told The Associated Press that the implications extended beyond Median itself.
“This is much bigger than one fake poll,” Wilson said.
“It’s never been easier to make invented numbers look legitimate, and campaigns and outside groups have every incentive to amplify anything favorable before anyone verifies it.”
That incentive is easy to understand. A poll showing a candidate unexpectedly dominating a primary can generate headlines, social media attention, donor enthusiasm and a perception of momentum. By the time questions are raised about methodology or credibility, the original numbers may already have circulated widely.
Median’s Wisconsin poll provided an unusually stark example.
A 23-point advantage is not a statistical squeaker that disappears because of ordinary polling error. It suggests a commanding lead. Hong’s eventual defeat made Median’s numbers impossible to ignore, prompting the organization’s initial admission that it had gotten the race wrong.
Only afterward did the larger story emerge: There had never been a legitimate survey to miss.
The Associated Press reported that Median’s website eventually went offline. Before disappearing, the site announced that the organization “has concluded its polling project and will not publish additional polls.”
Median says its purpose was to test whether polling information could spread “without independent verification.”
