A mysterious polling outfit that planted fabricated survey data into two major Democratic primary races has closed up shop after confessing the whole operation was a "social experiment", and nobody behind it will say who they are.
Median Strategies, a firm with no public track record and no named staff, admitted this week that polls it released in the Los Angeles mayoral race and the Wisconsin Democratic gubernatorial primary were invented from scratch. The company posted a statement on its website saying it had "concluded its polling project" and would publish no additional surveys. It withdrew all previously released data and told the Los Angeles Times it was "not seeking publicity or attribution for the individuals involved and are declining interviews."
The admission caps a weeks-long episode that exposed how easily bogus numbers can travel through the political ecosystem, picked up by local TV newscasts, shared by online outlets, amplified on social media, and even promoted by at least one of the candidates the fake data was designed to flatter.
In the Los Angeles race, Median Strategies released a poll claiming Mayor Karen Bass held a 12-point lead over City Council member Nithya Raman, a member of the Democratic Socialists of America. The survey cited responses from 560 Los Angeles voters and listed a 4.1% margin of error, details engineered to look legitimate. The Washington Examiner reported that the fabricated poll pegged Bass at 50% and Raman at 39%, complete with manufactured methodology and detailed breakdowns.
Bass herself shared the poll on X before it was debunked. She later deleted the post. Her campaign defended the move by saying the poll "was reported on by multiple news outlets", a defense that only underscores how fast unverified data can circulate when campaigns have every incentive to amplify favorable numbers and no incentive to question them.
When the June primary results came in, Bass edged Raman by five points, not twelve. Spencer Pratt, a Republican running as an independent, finished third. Bass and Raman now head to a general election.
After the hoax was exposed, a spokesperson for Bass demanded accountability. The New York Post reported the mayor's team called for the "full force of law" against those responsible:
"Any bad faith attempts to influence elections should be investigated and prosecuted to the fullest extent of the law."
That is a reasonable demand. But it sits awkwardly next to the fact that Bass's own account amplified the fabricated data before anyone bothered to verify it.
The Los Angeles survey was not an isolated stunt. Median Strategies also released a poll ahead of the Wisconsin Democratic gubernatorial primary showing state Representative Francesca Hong, another DSA member, leading Milwaukee County Executive David Crowley by 20 points. Breitbart reported the gap was listed at 23 points in the firm's release.
The actual result was a mirror image. Crowley defeated Hong by a razor-thin margin in the primary held the week of August 10. Hong arrived to vote in Madison on August 11; Crowley had greeted supporters at a Milwaukee rally the night before. The fabricated poll had the race backwards by more than two dozen points.
Median Strategies' own statement acknowledged the scope of the deception. It said all previously published polling releases "have been withdrawn and should not be cited or treated as genuine polling data." The firm told Fox News it was not affiliated with any candidate, political action committee, or party.
Nobody knows who ran Median Strategies. The firm declined interviews. It refused to identify the "individuals involved." It offered no registration records, no staff bios, no client list. The entire operation appeared, injected false data into two contested Democratic primaries, and vanished behind a self-congratulatory statement about studying the "political information ecosystem."
Calling it a "social experiment" does not make it one. Legitimate researchers publish methodology, submit to institutional review, and disclose their identities. Median Strategies did none of those things. What it did was fabricate election data, watch it spread, and then walk away.
Rick Taylor, a veteran political consultant, told the New York Post the damage goes beyond one race:
"Whoever is responsible for this ought to be exposed, and we need to make sure this never happens again."
Republican pollster Chris Wilson, who leads the firm EyesOver, framed the problem in broader terms. As AP News reported, Wilson warned that the episode reflects a systemic vulnerability:
"This is much bigger than one fake poll. 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."
Not everyone fell for it. AP News reported that major polling aggregators including The New York Times, Real Clear Politics, and FiftyPlusOne did not include the Median Strategies surveys in their databases because the firm's methodological information was missing. Mary Radcliffe, head of research at FiftyPlusOne, described the fragility of the system those gatekeepers try to protect:
"This is an industry built on trust. I have to trust that people are answering my questions in good faith."
But the gatekeepers are not the whole ecosystem. The fabricated polls still reached social media feeds, local television broadcasts, and online political reporting outlets, channels where millions of voters actually consume information. The fake numbers entered the bloodstream of two primary campaigns without anyone demanding proof that the firm behind them actually existed.
The Median Strategies hoax landed during a primary season already marked by significant polling misses in Democratic contests. In Michigan, polls widely inflated the lead of Abdul El-Sayed over Representative Haley Stevens in their Democratic primary. El-Sayed, a progressive candidate, ended up edging Stevens by only a narrow margin, far short of the comfortable lead surveys had projected.
In Minnesota, the error ran the other direction. Polls underestimated the strength of Lieutenant Governor Peggy Flanagan, a progressive, in her Democratic Senate primary against Representative Angie Craig. Flanagan captured the nomination by roughly 20 points, a margin the pre-election surveys failed to anticipate.
Whether those misses stemmed from methodological failures at legitimate firms or from something else remains an open question. But the pattern is clear: voters heading into contested Democratic primaries this cycle had access to polling data that repeatedly failed to reflect reality. Median Strategies simply took that dysfunction to its logical extreme by skipping the polling altogether and publishing numbers it made up.
The AP report noted that the fake Median Strategies poll in the Los Angeles race appeared to trigger suspicious trading activity on election prediction markets shortly after Bass shared it, a detail that raises questions about whether someone sought to profit from the fabricated data.
Several basic questions hang over this episode. Was Median Strategies registered as a legal entity, and if so, in which state? Did the fabricated polls measurably affect voter behavior or candidate strategy in either Los Angeles or Wisconsin? Are any law enforcement agencies or election regulators investigating? And were the Michigan and Minnesota polling failures connected to Median Strategies, or were they separate breakdowns by other firms?
Bass's team wants prosecution. Taylor wants exposure. Wilson wants the industry to reckon with how easy it has become to counterfeit credibility. All three demands point to the same problem: a system in which an anonymous entity can manufacture election data, watch it ripple through campaigns and newsrooms, and then dissolve itself with a press release, calling the whole thing an experiment, as if democracy were a lab exercise.
When anyone can invent a polling firm, fabricate results, and watch candidates and media amplify the fraud without a single phone call to verify it, the problem is not just one fake outfit. It is every institution that treated the numbers as real because the spreadsheet looked professional enough.
