In the days leading up to an election, I have two predictable habits. Each morning when I wake up, and occasionally when I doom scroll at 3 a.m., I obsessively check the polls. I go to FiveThirtyEight, then to the Washington Post, then check Quinnipiac. I look at them seeking some sort of certainty. If they don't give me the answers I like, I keep looking. I want them to tell me that the candidates I care about are going to win. When they don't, I keep checking, hoping that their predictions will shift. When they offer favorable results, I worry they will change. So, I check them again. By the time we are a few days from Election Day, I am checking them about 10 times a day.
I can also be counted on to never, ever answer a poll — whether the requests come to me via email, spam call or text. I am not doing it. I never have. Sometimes when I am out with friends we joke that none of us has ever done one. Who has time for that? Who picks up calls from unknown numbers? I have yet to find a single friend who tells me they have answered a poll. Even weirder, we seem pretty smug about the whole thing.
Even those of us who imagine ourselves to be politically savvy can be stunningly stupid.
Then, as if that weren't enough, when election results come in and they differ from the polls, and when this means a candidate I thought would win, doesn't, I am crushed. Like stuck on my couch in my PJs at 4 in the afternoon down. How could the polls have been wrong? I thought we had this.
I am an idiot.
This bifurcated relationship with polls has, up until now, made perfect sense to me. And yet, it reveals that even those of us who imagine ourselves to be politically savvy can be stunningly stupid.
So, let's start with what we know and see if we can stop being quite so dumb about polling.
First, polls are wrong. They are wrong all the time. They were wrong when they predicted Hillary Clinton would beat Donald Trump and they were wrong when they predicted the wins the Democrats would have in 2020.
Each time they are wrong, they tell us next time they will be less wrong — only to, at times, be even more wrong.
Second, answering polls doesn't appeal to everyone. Here I am not just referring to myself and my smug friends. In a story for Vox after the 2020 election, Dylan Matthews pointed out that the kind of people who answer polls are weird. In it, he interviews pollster David Shor, who explains that the type of person who answers a poll is generally quite different from one who doesn't, and that discrepancy means polling will inevitably be off.
"The reason why the polls are wrong," he explains, "is because the people who were answering these surveys were the wrong people."
The people who don't answer polls do look at them. And when they do, voting plans can change.
Shor also contends that the type of person who answers a poll tends to skew Democrat, which leads to overestimating the Democratic vote. People who answer polls are more politically engaged and have higher social trust, a trait that correlates with being a Democrat. Research has shown that conspiracy theorists, like followers of QAnon, aren't going to answer a poll.
But the breakdown isn't only across parties. Research also shows that millennials answer surveys far less frequently than older generations. (They came to this conclusion via a survey, though — how's that for a head-scratcher.)
Other factors include who answers phone calls from unknown numbers, who takes the time to do online surveys, and who will answer a random text. That's the bucket I and my friends are in. We aren't answering polls because we are already getting so much spam that when we see an unknown number or a random text or email, we just ignore or delete it. For us, it isn't about political engagement or social trust — it is about how we interact with the onslaught of unsolicited communication in our faces every day.
So, polls get it wrong because only weird people answer them. But what's even worse? The people who don't answer polls do look at them. And when they do, voting plans can change. Research shows polls depress turnout, especially if one's candidate is shown to have either a low or a high chance of winning. In fact, the only helpful polls for turnout are the ones that suggest that the outcome of the election is a toss-up.
Polls shouldn't be for voters, but rather for strategists. To the extent that polls help, when they are accurate, they can determine where energies and resources are best directed. They identify the close calls so a campaign can put more effort where it will matter. Polls can also help identify what is important to voters, but again, if respondents are skewed, it isn't clear how useful that may really be.
But that's the thing — polls aren't just consumed by those working on campaigns. They are often the dominant source of information used by news organizations covering elections.
When polls are used to help the news media frame its coverage of elections, the results are highly negative. News media coverage of polling, regardless of the accuracy of the polls, doesn't just lower voter turnout; it turns coverage of elections into a focus on the horse race, rather than the issues.
It turns democracy into a game of winners and losers, lowers turnout and is often wrong.
Horse race coverage, first of all, frames elections into sporting events. As Denise-Marie Ordway explains, "When journalists covering elections focus primarily on who's winning or losing instead of policy issues — what's known as horse race reporting — voters, candidates and the news industry itself suffer."
In a story on the failures of the news media during the 2016 election, Thomas E. Patterson, professor of government and the press at the Harvard Kennedy School of Government, underscored the fact that horse race coverage is entirely dependent on obsessing over polls. He notes that in the lead-up to Election Day 2016, "well over a hundred separate polls — more than a new poll each day — were reported."
Horse race reporting correlates with lower voter trust in politicians and news outlets as well as a less informed electorate. It turns democracy into a game of winners and losers, lowers turnout and is often wrong. Describing what has been called "The Nate Silver Effect," Benjamin Toff claims that news coverage of polling has eroded journalistic standards.
All of this is bad for a functioning democracy. Focusing on who will win takes attention away from why a candidate should or shouldn't win, what they stand for and what policies they may advocate. When polls suggest clear winners and losers, they eliminate the role citizens play in the process by suggesting their role as voters is unnecessary. Turning everything into a game that can be predicted eclipses the fact that democracies require ongoing engagement, public debate and responsive institutions.
So stop looking at the polls. Stop expecting them to offer clear and accurate predictions and stop being surprised when they don't. Definitely stop thinking that if you don't participate in them, they have a shot of being right. But most importantly, stop thinking that an imprecise set of statistics is going to save democracy, when you know that the only one who can do that is you.