A mysterious polling operation successfully deceived a major Democratic campaign, moved real money in election betting markets, and vanished almost as quickly as it appeared.
Los Angeles Mayor Karen Bass promoted a survey last week that purported to show her opening a comfortable lead in the city’s mayoral runoff against City Councilwoman Nithya Raman. However, the poll was fabricated.
The group behind it, Median Strategies, later admitted that its results were made up and described the entire operation as a short-term “social experiment.” The organization has refused to publicly identify who ran it or why three active American elections were selected as their laboratory.
This incident was not an internet prank that remained hidden. Bass’s reelection campaign amplified the fake numbers as evidence of her “gaining momentum” before deleting the post after the deception was exposed.
The operation involved at least three fabricated surveys covering races in California, Wisconsin, and Nevada. The Los Angeles poll claimed Bass had a nearly 12-point lead over Raman in the November runoff.
Median Strategies also falsely reported that Wisconsin Democrat Francesca Hong held a staggering 23-point primary lead over David Crowley. Hong lost, and the company soon admitted the survey was not real.
The fake polling firm presented itself as an independent entity with professional-looking graphics and a documented margin of error. It claimed to have surveyed 560 likely Los Angeles voters between July 30 and August 5, using methods it described as “mixed-mode interviews” and “transparent weighting.”
Despite its claims, the company declined to identify itself or answer basic questions from reporters. Once exposed, Median Strategies withdrew all polls, shut down its website, and insisted that no one involved had traded on election prediction markets.
The fraud did not remain confined to political discourse. Within minutes of Bass’s campaign sharing the fake numbers, betting markets reacted: The price of a contract predicting a Bass victory on Kalshi rose from roughly 63 cents to 65 cents within 15 minutes. Similarly, about 20 accounts traded thousands of contracts in her favor on Polymarket within six minutes.
The operation began posting online with only a tiny social media following. Established polling aggregators had spotted warning signs that Bass’s campaign apparently missed, as they did not include the surveys due to missing basic information about the firm’s ownership and methodology.
Anonymous individuals deliberately introduced fabricated polling data into active election processes before vanishing, having demonstrated how easily such misinformation could spread.