‘Shocking’ fake poll firm puts industry under the microscope: ‘Unfortunately not surprising’

The revelation that two polls released in recent weeks from a purported polling firm were faked is adding a new curveball for a polling industry already trying to improve public trust in its work.

Median Strategies made the stunning announcement earlier this past week that its poll results of the Wisconsin Democratic gubernatorial primary and the Los Angeles mayoral race were not real and instead were put out as part of a “social experiment.” It said they should not be cited or treated as legitimate data.

Coming off well-publicized polling misses in the Michigan Democratic Senate primary and the Wisconsin gubernatorial primary, the timing of the ploy wasn’t helpful either.

Pollsters said this incident emphasizes the due diligence that media outlets and the public must conduct when looking at and trying to understand numbers being released.

“It was very shocking, but unfortunately not surprising,” Ben Leff, the CEO of the polling and survey research firm Verasight, said about Median’s fake polls. “I think the industry and media are used to probably polls that aren’t conducted very well in terms of rigorous methods or how they’re verifying respondents, but this was sort of the extreme of that, where everything was completely fake.”

The polling industry has already been dealing for years with struggles to improve its credibility in the eyes of the public, with trust shaken by President Trump’s repeated outperformances in 2016, 2020 and 2024, to varying extents .

But that has been at least based on good-faith efforts from pollsters trying to model what an election may look like the best they can. That’s a different situation from Median, which pollsters emphasized shouldn’t be considered part of the industry.

Median first released a poll on the Los Angeles mayor’s race, a contest that hasn’t seen any major public polling conducted since the primary election in June. It showed Mayor Karen Bass (D) leading progressive City Council member Nithya Raman (D) by 12 points, a lead that Bass touted in a later-deleted post on the social platform X.

After Median revealed the poll was fake, Bass’s campaign denounced it, with a spokesperson calling for any “bad faith attempts to influence elections” to be “investigated and prosecuted to the fullest extent of the law.”

Raman’s campaign attacked Bass for promoting a “fake poll without properly vetting the source.”

A second fake poll from Median showed Wisconsin state Rep. Francesca Hong leading the state’s Democratic primary for governor by 20 points. The democratic socialist ultimately lost the primary earlier this month in a major upset .

The poll’s result was largely in line with what several other polls from legitimate firms were indicating, though polling was limited in the primary race’s final weeks.

The extent to which the poll influenced expectations, and potentially turnout in the race, is impossible to quantify, but the more immediate impact may be how polls are perceived as the country is gearing up for the critical midterm elections in November.

Geoffrey Skelley, the chief elections analyst at Decision Desk HQ, noted the impact is different than if a major pollster was revealed to be falsifying data. But he said he understands how this could affect perception.

“I understand why, say in the aftermath of 2016, people have that attitude,” he said. “That is concerning, that it sort of adds fuel to the fire of ‘The polls are worthless,’ ‘The polls are meaningless,’ ‘They’re not even real’-kind of conversation.”

Median maintained on its website that it wanted to see what would happen if it put out fake polls.

“Median Strategies was created as a short-term social experiment examining how purported polling information could enter and spread through the political information ecosystem without independent verification,” its message read.

Median did not return a request from The Hill for comment.

The Guardian identified the source behind Median as 21-year-old Rahil Prakash, a recent college graduate. He said he was inspired by the polling miss in the Michigan Senate primary to see if fake polls could “penetrate the ecosystem” easily.

“I was a bit shocked, perhaps at how fast this sort of information can spread,” he said.

Prakash said he wanted to apologize to the Bass, Raman and Hong campaigns.

“And I hope it gives a really good lesson going forward in that we need to be very serious about these things and very serious in how, if one person can do this, what can a serious operation do in the future,” he said.

Experts expressed skepticism about Median’s claims about its motive, noting various political or financial incentives for the results. The rise of prediction markets in recent years gives another potential impact on this, as bettors place wagers on a wide range of questions, including on who will win an election. Prakash told The Guardian he placed no bets on the races.

Skelley said fake pollsters have put out results in the past intended to influence prediction markets and make money off the impacts. But they have generally not come off looking as legitimate as Median did, either not having a website or having a suspicious-looking one.

Leff said the first step that media outlets and members of the public should do if they see a poll from a pollster they don’t recognize is check if the pollster is a member of the American Association for Public Opinion Research’s Transparency Initiative .

He said the initiative is similar to the bar for attorneys and provides credentials to pollsters that are transparent about disclosing their methodologies in detail.

He said a second step is looking for a detailed methodology statement provided to the public included in the poll’s release. That statement would include information such as the number of respondents, margin of error and how respondents were recruited.

Leff said Median had a “top-line” report of its methodology that didn’t provide a lot of details.

“That was immediately a red flag to us,” he said.

Of course, even a detailed methodology statement could still be faked, and conducting a deep dive may be more extensive than the average person knows to do or takes the time for.

John Cluverius, the director of survey research for the UMass Lowell Center for Public Opinion, said this underscores the importance of not giving too much emphasis to any single poll but what pictures various polls from different pollsters are painting.

“Thinking about polls in terms of them being different snapshots that are trying to measure the same thing, I think that’s very useful,” he said.

Cluverius said he believes media outlets have become “smarter” about insisting on transparency from polling firms, becoming more selective in which polls they cover and don’t, which can help weed out those that are faulty.

Leff noted this is just one of the threats the polling industry faces, pointing to bots and artificial intelligence that attempt to take surveys.

In another incident, the polling aggregator FiftyPlusOne announced Thursday that it would “permanently ban” one polling firm and suspend another over a lack of disclosure that Florida GOP gubernatorial candidate James Fishback’s campaign was funding them. The banned firm was also accused of fabricating data, a claim that it denies.

Ryan O’Donnell, the executive director of the left-leaning pollster Data for Progress, said the takeaway from the controversy needs to be more vigilance.

“The lesson here isn’t that we should stop trusting polling — understanding public opinion is vital to a functioning democracy and there are plenty of credible, trustworthy pollsters out there,” he said. “People just need to know what to look for.”