The Pollsters’ Epic Fail: 2020 is Not a Technical Problem

By Kent M. Harrington

Missing their calls on two presidential elections have the pollsters scrambling to fix what went wrong. They need to consider whether their assumptions about the Americans they’re surveying make sense anymore.

Chief Bromden, the broom-pushing patient who narrates One Flew Over the Cuckoo’s Nest, Ken Kesey’s novel about the never-to-be forgotten characters confined on a psychiatric ward, nailed what the polls forecast in the 2020 election. “It’s the truth,” he said, “even if it didn’t happen.” Like toxic household products that must display warning labels, the pollsters who blew their second, consecutive presidential contest big time should have the Chief’s line tattooed on a prominent body part. It’s a buyer-beware notice of more failures to come.

The pollsters are picking through their algorithms and analysis to figure out what went wrong. The fact that their statistically sanctified predictions of red mirages and blue waves cratered is non-trivial as well as embarrassing. Unfortunately, as in 2016, the profession is producing technical explanations, rather than serious thought. Like the doctor who emerges from the operating room to talk to the waiting family, it’s the polling business version of the surgeon’s one liner—the operation was a success, but the patient died.

To be sure, pollsters have real problems and are acknowledging as much. Most obviously, the people they needed for their cross section of voters are far harder to reach in 2020 than in, say, 1976. No more landlines at fixed addresses, English-speakers at the end of every call, or even polite phone etiquette. Instead, with smart phones that the law says can’t be autodialed, caller ID that makes pollsters easy to ignore, and Millennials and Gen Z’s who would rather text than talk, the good old days of easy outreach to the man-in-the-street are long gone.

Some pollsters also point out voters are making it difficult to pick their brains. Their rationalizations vary for who they missed and why. Many cite voters who won’t talk about their choices. Others say younger people are elusive; like rural Americans and Hispanics, their behavior means they weren’t sampled in sufficient numbers to get a good read on what they think. Worse, many first-time voters in 2020 mulled their choices until election day. The bottom line: a turbulent as well as truculent electorate didn’t play the opinion surveyors’ game.

Not surprisingly, the pollsters are facing fire on several fronts. Campaigns that paid for missed predictions aren’t happy. Neither is the commentariat that is making clear that pollsters led them astray. Critics also are reminding that none of the pollsters’ technical problems are new. Others more charitably are waxing philosophic, recalling that political prognostication is an art as well as a science. It’s all true, of course. But none of the critiques address the big picture: an American landscape that may well have left political polling permanently at sea.

Consider history and the polls that foresaw a blue wave. Whether it’s generations of voters who have chosen divided government—the President’s party has held both house of Congress in only eight of the last 50 years—or stagnant Republican and Democratic party memberships, since the 1960s most elections have brought changes in the political balance that are ripples, not ground swells. 2020 is a case in point. Four years of Donald Trump’s failed policies and dictatorial pronouncements galvanized Democrats. But they didn’t budge 70 million Republicans and GOP-leaning voters who backed him on November 3rd.

Appearing to ignore a half-century of political history isn’t the only thing that raises questions about the pollsters’ election day calls. Their own surveys also make clear that today’s voters are cementing these political divisions in place. A Pew Research Center poll reported this month that 90 percent of Democrats as well as Republicans saw the other’s presidential candidate as a looming disaster. The issue wasn’t just Biden or Trump; 80 percent in both camps said they disagreed with the other’s “core values and goals.” In October, 85 percent of Americans queried told Pew the two sides couldn’t even agree on basic facts.

Combine history with hardening attitudes and the question is obvious: how much can election polls really predict if they presume voters see choices on the ballot when the vast majority don’t? It’s not a trick question. Pollsters have collected data on the country’s fractures and assessed their significance for years. But the polling gurus have said nothing about their effects on behavior at the ballot box, not to mention on answers to their own opinion surveys, when voters don’t believe they have a real choice.

Take Hong Kong. It’s an example of what pollsters are ignoring here. Until this year, its citizens could vote for a semi-democratic legislature and enjoy free speech and assembly, all widely supported rights according to countless surveys. Today, Beijing is imposing draconian national security controls that strip away these rights and polls are painting a different picture. Support for democracy remains strong, but not for protests against China’s repression. After all, when voicing support for protesters can put you in a neighboring prison cell, what’s your choice when pollsters call?

To be sure, Hong Kong citizens criticizing an authoritarian regime is one thing, Americans expressing their political views in an overheated election is another. But even as pollsters are chronicling the shift in how voters see each other and their choices in the United States, they seem to be missing the point and perhaps the issue important to their own success. Are the effectiveness and value of opinion surveys in portraying the American political landscape diminishing because of its disappearing comity and common ground?

Finding an answer obviously is complicated. It’s not just the voters’ fault. From social media’s malignant technologies marketed as campaign tools to the cable networks’ micro-targeting of partisan political ads to amplify their effects, the media business is happily peddling wares that are widening the country’s political divide. Add the news deserts created by the demise of print journalism. Across the country, much of the once rich, local reporting that provided context for polling’s statistical depictions has disappeared.

As they search for answers on how to do better next time, the pollsters should mull Neil Postman’s insights. The late media and cultural critic observed that new technologies always compete with old ones for time, attention, and for the dominance of their world view. A new technology, Postman wrote in 1992, doesn’t add or subtract. It changes everything.

In a deeply divided America, Postman’s observation should be encouragement for the polling industry to check its assumptions. Like the guy in the Hong Kong high rise when the pollster calls, for many Americans the problem just might be that it doesn’t make sense to speak your mind anymore.

Appeared first in MediaVillage