Median Strategies said its median strategies bogus poll la was not a real poll, after a Los Angeles mayoral survey showed Karen Bass ahead of Nithya Raman by nearly 12 percentage points. The company said the fake result had been built to test how polling claims can spread before anyone checks them.
Karen Bass and Nithya Raman
The false survey said it reached 560 Los Angeles voters between July 30 and Aug. 5 and carried a 4.1% margin of error. It also said 16% of respondents had voted for Spencer Pratt, with about 85% of those voters backing Bass and about 7% backing Raman.
Median Strategies described the project in an email to The Times as a short term social experiment to examine how purported polling information could enter and spread through the political information ecosystem without independent verification. It also wrote that most Pratt voters appeared to be treating Bass as the more familiar, more moderate option rather than sitting out the race.
Bass campaign on X
Last week, Bass posted on X from her campaign account, “Doing the work, showing up, and gaining momentum. Let’s do this, LA!” The Bass campaign later said, “This poll was reported on by a news outlet,” and added, “Any bad faith attempts to influence elections should be investigated and prosecuted to the fullest extent of the law.”
That sequence is what makes the bogus poll more than a stray online mistake. Bass publicly used the number before Median Strategies said it was fake, while Controller Kenneth Mejia separately pointed on X to polls that had shown him losing before he won by a wide margin.
Median Strategies and Wisconsin
Median Strategies began posting on social media on Aug. 9 and had about 20 followers. On Aug. 12, it wrote, “We got Wisconsin wrong- and by quite a margin too,” after releasing a Wisconsin gubernatorial poll showing Francesca Hong up by more than 20 percentage points.
For readers trying to judge the next round of political numbers, the practical lesson is simple: a poll that looks precise can still be a test run, not a measurement. In this case, the company itself said the point was to show how unverified polling can move through the information stream before anyone asks who made it, when it was collected, or whether the numbers can be trusted.
