They are from different worlds, different parties, different everything, but Florida’s Democratic gubernatorial candidate Andrew Gillum and President Trump have at least one thing in common: Pollster after pollster failed to see their surprising victories in the days before Election Day or gauge the anti-establishment nerve both men struck with voters.
Gillum, the African-American mayor of Tallahassee, made history when he came from behind on Tuesday to beat all of the Democrats considered the front-runners in the gubernatorial primary. He defeated Gwen Graham, the expected winner. The thinking went that former Miami Beach Mayor Philip Levine would come in second, with Gillum behind them in third place, possibly fourth. That’s what pollsters predicted, and that’s what the media relayed to readers.
But there was a critical disconnect.
In the light of day, it’s clear that the way voters’ intent is measured can be faulty, creating misleading information that gives pundits and politicians talking points, but offers little of substance to voters.
Gillum won with 24 percent of the vote and will face Republican Ron DeSantis in November.
Why are pollsters so off the mark when it comes to predicting some major races? It’s happened in other parts of the country, as in New York with Alexandra Ocasio-Cortez’s congressional race.
Pollsters appear to be baffled, too. No Florida poll had Gillum winning in Florida on Tuesday. Gillum didn’t even have TV ads running until the final week of campaigning. He didn’t lead in a single major poll and was behind in fundraising. Now significant in retrospect, however, is the influx of millions to the Gillum campaign from liberal billionaires Tom Steyer and George Soros, the endorsement of Sen. Bernie Sanders and a grassroots “Souls to the Polls” campaign among African-American voters in the last days of early voting. But the establishment failed to believe Gillum when he kept assuring supporters and reporters that he was “surging.” The claim was ignored — because of the polls.
Brad Coker, managing director of Mason-Dixon Polling & Strategy in Jacksonville, says it’s an example of how “a progressive Democrat” can fly low under the radar” until Election Day. “Gillum was pretty stealthy,” Coker told the Editorial Board. He theorizes that Sanders voters and black voters who backed Hillary Clinton in 2016 came together to vote for Gillum, giving him the victory.
Doug Kaplan, president of Galvis Marketing, which gauged the governor’s race for months, told the Board that a key might have been that the predicted one-third of undecided votes went for Gillum.
“Graham and Levine crashed in South Florida,” he told the Board in an email. “Graham fell apart with Election Day voters,” he said. On top of that, turnout increased significantly on the Democratic side — adding unmeasured voters.
In addition, many polling firms traditionally query “likely voters,” those with landlines to call. They’re going to have to pivot, connecting with people who are the targets of untraditional campaigns. For instance, social media clearly reach a different demographic, younger, perhaps less drawn to establishment candidates.
If there’s one takeaway, it should be that polls can be inexact and, sometimes, unreliable. It’s good to see voters betting on their own instincts, rather than on the horse races polls too often create.