Pollsters face reckoning after massive errors in presidential, congressional races

Political pollsters emerged among the biggest losers of election night 2020 after incorrectly forecasting massive wins for congressional Democrats and Democratic presidential nominee Joe Biden.

Unbent by reputational beatings for failing to predict President Trump’s 2016 victory, pollsters had Biden coasting to the White House with an electoral and popular vote landslide, with potential wins even in Republican strongholds Texas and Ohio.

In Florida, polling averages had Biden leading before Tuesday. Quinnipiac University, a well-established pollster, had Biden ahead five points in an Oct. 28-Nov. 1 poll. But Trump carried Florida by an easy 3.4 points with roughly 96 percent of ballots counted.

In Texas and Ohio, Democrats pined for shocking upsets, backed up by polls showing Trump tied or even losing. Polling aggregators including RealClearPolitics rated both states as toss-ups.

In Texas, Quinnipiac and Emerson College had the presidential candidates tied in the final stretch. In Ohio, Quinnipiac and Emerson had Biden ahead, though Republican-affiliated Trafalgar Group and Rasmussen Reports had Trump up by four or five points.

The president trounced Biden by six points in Texas and eight points in Ohio as of Wednesday afternoon, with most votes counted.

Across Midwest battlegrounds, most professional pollsters had Biden vastly ahead on Election Day, despite very close races as absentee ballots are counted on Wednesday.

In Wisconsin, where Biden leads by a razor-thin 20,000 votes as of Wednesday afternoon, with a recount possible, the RealClearPolitics average gave Biden a 6.7-point lead. A recent CNN poll had Biden up eight points, the New York Times had Biden winning Wisconsin by 11 points and an Oct. 20-25 ABC News/Washington Post poll had Biden crushing Trump by 17 percentage points.

Trump routinely claimed during campaign rallies that the polling industry again was failing to accurately count his supporters, some of whom, he said, were less inclined to publicly state their support.

As the closeness of races in 2016 battlegrounds became clear, Democratic activist and CNN contributor Van Jones bluntly declared, “These polls are not to be trusted. There is something wrong with the polling industry.”

And polling errors in favor of Democrats impacted not only Trump, whose race against Biden remains undeclared due to narrowly divided swing states. In House and Senate races, pollsters also incorrectly gave Democrats an advantage across the nation, inspiring liberals to pour millions into races that ultimately weren’t close.

In Maine, incumbent Republican Sen. Susan Collins defeated Democrat Sara Gideon, with about 51 percent to Gideon’s 42.4 percent, with an estimated 85 percent of ballots counted, according to the Associated Press. But Gideon led Collins in more than a dozen recent public polls and trailed Collins in none of them.

Gideon, who conceded defeat, was up by six points in a poll conducted Oct. 29-31 by Emerson College and was ahead by five points in mid-September surveys by both the Boston Globe and the Times.

“You can’t look at some of the presidential or Senate polls (Florida! Maine! Michigan!) and think pollsters are doing well. Overall House predictions were off 15-25 seats. A disaster,” wrote Politico’s congressional bureau chief John Bresnahan.

“Systematic failure,” Fox News pollster Frank Luntz wrote about the Maine Senate race.

Michigan Republican Senate candidate John James, meanwhile, led Democratic incumbent Sen. Gary Peters by almost 10,000 votes out of the more than 5.1 million counted as of Wednesday afternoon, with the AP saying 96 percent of results were in. James was expected to lose by a sizable margin, according to every recent public poll. A Reuters/Ipsos survey from Oct. 27 to Nov. 1 had James down seven points.

“The achilles heel of the polling industry: a lack of humility, a smartest-guy-in-the-room attitude,” wrote Politico journalist Tim Alberta, circulating tweets from a data journalist for The Economist, who accused him of credulously believing Michigan would be close.

“This ‘expert’ spent a day mocking my reporting (and [Michigan Democratic Rep.] Elissa Slotkin’s belief) that Trump voters were being undercounted in polls of MI and her district. Which — surprise! — they were.”

Share this article:

Source: Read Full Article