Commentary: Predictive political polling is a time-honored part of political news coverage. Trouble is, it’s a very broken tool.
Given the chasms of competence and character that separate the two presidential candidates, it’s hard to imagine that a large number of people still have not decided whether Kamala Harris or Donald Trump deserves their vote.
If you believe the polls, likely voters (those mainly partisan stalwarts who take elections seriously) are nearly evenly split between Harris and Trump.
But an estimated one-sixth of the U.S. electorate is for one reason or another uncommitted, according to a recent New York Times/Siena College poll. They’re still making up their minds — both on the national level and in the seven swing states whose Electoral College votes will essentially decide the presidential election.
There has been a strong push in recent years, led by media critic Jay Rosen, for news organizations to focus more on the issue-driven stakes of an election, rather than on the statistical odds. And while there’s been lots of solid issue reporting in 2024, once again the gamification of politics is in full swing. No surprise here: A binary is such an easy story to tell — winners and losers, messy human concerns hammered down to data points, percentages, charts and graphs.
Polls are part of every major news organization’s political coverage. Trouble is, predictive political polling is a deeply flawed tool.
Our world has changed. It’s getting harder to assemble statistically valid samples on which pollsters base their predictions. [Ref. Pew Research Center] People are more reluctant to express political views. Noncompliance rates are soaring for phone surveys: Caller ID means fewer people answer when a stranger calls. When they do pick up, not everyone tells the truth. Polling samples skew white, partisan, middle class and old.
Opt-in online polls showed early promise, but over time they’ve been proven methodologically weak, harvesting poor quality data and projections that miss the mark. Most polls overestimated Joe Biden’s lead in 2020. And the “red wave” predicted in the 2022 midterms never came ashore.
The current tsunami of polling data is overwhelming and often unreliable. Data points are presented and perceived as fact, instead of the soft and fungible numbers that they really are. Even so, polls can sometimes yield useful insights on voters’ dilemmas, feelings, issues and needs.
Easiest to identify are the swing voters: partisans who have become disenchanted by candidates and are willing to switch sides.
The Never Trump site The Bulwark, is a conservative news aggregator, an offshoot of the now defunct Weekly Standard. It provides a good look inside the mind of Trump-to-Harris swing voters like the Bulwark’s Tim Miller, a former GOP operative now firmly in the Harris camp. It’s also a good example of how YouTube has become a free-to-all source of election news, from major media organizations to online influencers like Miller.
Max Tani of Semafor recently reported that since Joe Biden dropped out of the race, Bulwark videos have drawn 88 million views, making it one of the most visited news sites on YouTube. Bari Weiss’s Free Press offers a parallel YouTube refuge for Democratic voters who swing the other way: they still revile Trump, but are disenchanted with their party on DEI, Israel and other issues.
Undecided voters are a more nuanced electoral force — more conflicted and impossible to predict. These are people pulled in opposite directions — repelled, perhaps, by Trump’s moral and legal failures, but drawn to his stance on immigration or the economy. Or drawn to Harris’ politics of joy, but unconvinced of her foreign policy chops. Can they sort it out? Will they show up to vote? We shall see.
Less explored by pollsters and perhaps more worthy of campaigns’ attention are disengaged or low-information voters.
YouGov, one of the last remaining reputable online polling organizations, released a study of disengaged voters just before Biden dropped out. It presents these voters as younger, primarily female, earning lower incomes and without college degrees. They pay little attention to politics, get their news from friends and social media and may have limited understanding of how government works.
Largely politically independent, these voters are skeptical about the claims of politicians; they are less likely to hold extreme views. Their top concerns are high prices and inflation. They worry that the country is going in the wrong direction.
You see versions of these voters at any grocery store, pushing a cart, keeping their kids in line, feeling inflation’s pinch in the checkout lane. They’re likely working a job — or two. Courting exhaustion, keeping afloat. Little wonder they’re statistically less likely to show up on Election Day than people who pay close attention to the news. But that doesn’t necessarily mean they don’t care.
Disengaged voters have historically leaned pro-Trump, but after his felony conviction, his favorability among them slumped. How these voters feel about Kamala Harris is so far unclear: they’re still making up their minds.
Maybe someone will conduct a poll.
I’d be interested to know.
— By Joan Connell
Ed.: The illustration for this commentary was created by a human being who prompted Google Gemini AI to “generate a gender-neutral image, in the style of an Edward Hopper painting — with an autumnal color scheme — of a person who can’t make up their mind.”
• • •
We welcome letters to the editor responding to or amplifying subjects addressed in Community Voices. If you wish to contribute to Community Voices, please send an email with a subject proposal to Executive Editor Mike Sato (msato@rockisland.com) and he will respond with guidelines.
Help keep the local news flowing — support nonpartisan, fact-based, no-paywall local journalism
with a donation to the Salish Current — news for people, not for profit.