For the second time in two months, a polling organization embraced by media outlets has admitted to fabricating its results. This time, the deception lasted much longer, fooled far more organizations, and appears to have inflicted greater electoral damage.
In August, Median Strategies admitted to making up polls that showed socialist Francesca Hong leading Wisconsin’s Democratic gubernatorial primary by more than 20 percentage points and Los Angeles Mayor Karen Bass leading challenger Nithya Raman by nearly 12 points. Hong instead lost to David Crowley by 0.5 points.
Caught by the Los Angeles Times, Median described itself as a “short-term social experiment,” supposedly examining how unverified polling could spread through the political information ecosystem. There is little evidence that its fabrications changed Wisconsin’s outcome. Los Angeles has yet to hold its runoff, and the poll’s influence there remains unproven. Disturbing as the deception was, its reach appeared limited.
Then came Tuesday’s admission from Wedgewood Polls. This deception was far more extensive, reaching national polling aggregators and a Michigan Senate primary in which one candidate says the invented numbers did much to destroy her campaign.
Wedgewood’s June numbers put Abdul el Sayed at 46%, Rep. Haley Stevens (D-MI) at 39%, and state Sen. Mallory McMorrow (D-MI) at just 9%. Its purported general election survey showed el Sayed beating Republican Mike Rogers, 49% to 44%, while McMorrow trailed. The message was unmistakable: Progressive el Sayed could win, and McMorrow was wasting everyone’s time. El Sayed campaign used Wedgewood’s numbers in a fundraising appeal.
“Donations dried up,” McMorrow wrote. “Volunteers were devastated and pleading for guidance on what to do. Reporters hounded me with calls to drop out.”
She withdrew July 5.
El Sayed, perhaps buoyed by some of McMorrow’s marooned supporters, then beat Stevens by less than 1 percentage point, a substantially narrower margin than Wedgewood’s 7-point June lead. Nobody can reconstruct an election without the misinformation, but a shift of less than 1 percentage point could have produced a different nominee. Dismissing the fraud as inconsequential requires assuming away precisely the effects McMorrow described.
The fabrications extended well beyond Michigan, showing Democrats James Talarico leading in Texas, Sherrod Brown leading in Ohio, Adam Hamilton leading in Kansas, and Troy Jackson leading in Maine. In Mississippi, Wedgewood put Democrat Scott Colom just 4 points behind Sen. Cindy Hyde-Smith (R-MS), fueling headlines about a possible historic upset. These invented results supplied campaigns and journalists with compelling stories about Democratic strength in Republican territory. These news stories undoubtedly affect public opinion — that’s what news stories do.
Major polling aggregators helped give those fabrications credibility. The New York Times, Nate Silver’s Silver Bulletin, and FiftyPlusOne, founded by former FiveThirtyEight journalists, incorporated Wedgewood’s fabricated polls into their polling operations. RealClearPolitics, however, avoided the fraudulent numbers — congratulations to that standout organization.
Like Median, Wedgewood portrayed its deception as an experiment exposing industry weaknesses. It claimed no intention of influencing an election. Conveniently, this explanation arrived after investigators confronted it. Calling fraud an experiment does not make its victims imaginary.
Polling helps allocate political resources. A candidate supposedly surging receives free coverage, donations, endorsements, and volunteers. A candidate supposedly collapsing loses those same assets. Fabricated polling manufactures momentum for one campaign while persuading another’s supporters to surrender.
The details of Wedgewood’s funding and ideological slant are unknown, but there is sound advice for anyone seeking to know who is behind a secret operation: Ask who gains. Democrats, especially left-wing Democrats such as el Sayed, gained from the false polling. We expect to find out that left-wing Democrats are behind this latest example of cynical political manipulation. You know, members of the party that boasts it’s saving democracy.
Media organizations must demand identifiable researchers, transparent methods, and evidence that interviews occurred before publishing numbers from unfamiliar firms. Aggregators must recognize that inclusion functions as an endorsement, however many methodological disclaimers accompany it.
Voters and donors should exercise skepticism. A professional website and a spreadsheet establish nothing. Choose candidates for their judgment, principles, and records. Treat polls as fallible evidence, especially when the organization producing them has no record.