The worst mainstream media habit: Distorting polls for clicks
By now, my readers know full well what I think of national polls taken nearly seven months before the election: They are worse than meaningless. Pervasive polling obsession winds up misinforming (and freaking out) voters while crowding out the essential aspects of a historic campaign.
For starters, constant polling hype frames the election as a horse race, devoid of moral or policy outcomes. Premature polling distracts us from what is critical and central — fourtimes indicted former president Donald Trump’s unprecedented attack on democracy. (When you want to “suspend” the Constitution, use a putsch to overturn the will of the voters, unleash the military on civilians and weaponize the Justice Department, you are pining to undermine democracy and turn America into something resembling Viktor Orban’s Hungary.)
At its worst, coverage of polling misleads voters. Consider the opening of a recent New York Times article. “President Biden has nearly erased Donald J. Trump’s early polling advantage, amid signs that the Democratic base has begun to coalesce behind the president despite lingering doubts about the direction of the country, the economy and his age, according to a new survey by the New York Times and Siena College,” writes Shane Goldmacher.
This is false. The previous poll was well within the margin of error; the poll Goldmacher is hawking is within the margin of error. There has been no statistical change. The narrative of “Biden behind! Now, he’s catching up!” is mathematically inaccurate. And yet, the poll sets up a narrative and frames coverage — Why has Trump lost his lead?
Can Biden sustain his comeback? — all based on a false premise.
Even if outlets presented polling honestly (i.e., “The poll shows the race is now statistically tied but our poll does not predict the future”), the obsessive focus on polling as if it were the news itself would distort perception of the race for a simple reason: We cannot know in April who will win in November.
This realization will shock those grasping at polls for emotional solace during a hair-raising election season. However, too many people who will vote in seven months have yet to pay attention to the race. Moreover, too many events are still to happen between now and November to enable us to say the race then will look the way the polls do now. An infinite number of variables (e.g., Trump could be convicted, Trump could be acquitted, a wider Middle East war could break out, Biden could negotiate a Middle East peace deal) are at play. Just think how a single speech, Biden’s State of the Union, affected voters’ perception. Polling before that event that didn’t account for it appeared even less relevant in retrospect.
Poll defenders will say, “The poll just shows what would happen if the election were today!” But the election is not today — and everything between now and then will determine the outcome. So what’s the point of polling in April? Honestly, it is a cheaply obtained, lazy way to generate an audience.
Early polling rarely reflects the outcome. Too much (it’s called “the campaign”) happens in the interim. That was true in 2020, 2016, 2012 and 2008. Pollsters struggle with extremely low response rates and uncertainty in determining who will vote months from now. But the problem is more basic.
Brian Klaas, a political scientist and author of the fascinating book “Fluke: Chance, Chaos, and Why Everything We Do Matters,” told an interviewer in January: “I have no idea who’s going to win the 2024 election. It’s impossible to say, and that’s because the world is going to change drastically in the next 10 months or so.” He adds that “anyone who tells you, ‘Oh, I know exactly what’s going on,’ they’re just lying to you.” Maybe they are just desperately trying to sell newspapers, draw eyeballs and attract clicks.
Klaas continued, “We don’t know what’s going to happen. I wish that there was a bit more humility in forecasting and a bit more recognition that chaos theory tells us that the ability to predict the future is a pipe dream. We’re not going to be in that world where we can imagine the future quite clearly.”
Psychologists tell us we have an inherent bias (specifically, projection bias) that leads us to assume that the future (Biden loses!) will resemble the present (Biden behind!). And yet, in April 2008, few anticipated the financial collapse that essentially doomed John Mccain’s campaign; in April 2016, no one imagined that former FBI director James B. Comey would pipe up with “new evidence” about Hillary Clinton’s emails 11 days before the election. And the “red wave” forecasters who spent months predicting a Democratic wipeout in 2022 never accounted for Dobbs v. Jackson Women’s Health Organization. In other words, stuff happens.
Worse, polling frenzy crowds out real news and vital questions that will shape our collective future. Why does Trump routinely slur his words and sound incoherent? How does Biden’s economic record stack up against the past few presidents? What has happened to maternal and infant mortality and poverty in states with forced birth laws?
Rather than pretend that political gurus can anticipate results (which in turn, creates media coverage to match expectations), the media would do better to focus on the “not the odds, but the stakes,” as New York University journalism professor Jay Rosen puts it. Not only are the stakes — the fate of democracy — immensely important in 2024, but the odds (the polls) at this stage are virtually meaningless. News consumers should evaluate coverage accordingly.
Psychologists tell us we have an inherent bias (specifically, projection bias) that leads us to assume that the future (Biden loses!) will resemble the present (Biden behind!).