This Is The Post About Being Wary Of The Exit Polls

Do not allow yourself to be misled!

While every presidential election cycle is unique in its own way, some discussions are perennial. How does a “caucus” work? What happens in the case of an electoral college tie? We’ll be writing these every four years until the heat death of the universe.

And this is the thing we write warning you to be wary of whatever preliminary results you see bubbling up as Election Day unfolds. You should be even more wary of this year’s new development: The election day projections some groups are producing.

The best strategy: Be wary of anything that says it’s projecting the race before the polls are closed. Don’t go looking for it. Don’t put any special provenance in any data you do see. Especially be wary of anything you see creeping up on your Facebook feed. Facebook has been a font of misinformation this election cycle. Some fake news site is inevitably going to circulate fake data that pretends to know who’s winning. Don’t trust it. If you want our advice, stay off Facebook altogether on election night. 


So right about now, you might be wondering why we even do exit polls, seeing as how we’ve likened them to some sort of hidden danger, like a nest of copperheads in your root cellar. Exit polls are actually very useful, informative tools, when used correctly. Early guesses of who will win aren’t correct uses.  

Every year, a consortium of media organizations (basically, The Associated Press news agency and the major network and cable news outlets) known as the National Election Pool conducts exit polls. This group dispatches interviewers to polling places all around the nation. They’re assigned the task of corralling certain voters as they leave their polling places to ask them to fill out a survey in which they account for whom they voted and to compile a raft of demographic information.

As you might surmise, the exit polls yield a familiar product for your election night news coverage ― the slicing and dicing of different demographics to reveal smaller trends in the larger electorate. When CNN’s John King reports how single white women are voting in the Philadelphia suburbs, that data comes from exit polls. When MSNBC’s Steve Kornacki reports that national security was the key issue of voters in Florida’s I-10 corridor, he’s pulling that information from exit polls.

The exit polls also help the AP and the news networks ― traditionally tasked with the responsibility of “calling the race” in the various states, set their expectations for how the night is going to go and how likely it is that the call is going to bend one way or the other.

Exit polls are no small operation. It’s not something that you throw together by saying, “Hey guys, let’s do an exit poll. Here, print up some badges.” Edison Research, the firm that conducts the exit polls, works all year to prepare for this. They determine which precincts will be sampled (not all polling places get polled ― it’s a sample, just like with all polls), coordinate and train workers to randomly select respondents and conduct the poll and coordinate communicating the information back to the central database to be processed.

Results from these official NEP exit polls are kept secret most of the day. The first releases were due to come early Tuesday evening, but ― and we can’t emphasize this enough ― they should not be used to project who’s winning, who might triumph or anything about Hillary Clinton versus Donald Trump.


At some point during the night, the exit poll Clinton versus Trump  numbers might leak. And if they do, they are going to look very much like straight up state-by-state election results. And this will either touch off some premature bedwetting or way-too-early celebrating. If you were a big fan of John Kerry in 2004, you probably still carry the psychological scars of an evening that began looking like a certain Democratic victory to an election that eventually wound its way round to the reelection of former President George W. Bush. And let’s not even talk about 2000.

More recently, the early exit polls in the 2012 Wisconsin recall election briefly fueled Democrats’ hopes that they were on the verge of victory. Early exits polls indicated a dead heat between Wisconsin Gov. Scott Walker and his opponent, Milwaukee Mayor Tom Barrett. But as The Washington Post’s Jon Cohen reported, “Just a half hour later, the exit poll shifted to 52 to 48 percent, tilting in Walker’s favor. (The final margin appears to be seven percentage points.)”

Here’s a rundown of everything that makes the exit polls imperfect:

1. Remember that the exit polling data is going to first show up in network and cable news coverage. The data will be used to fill the long gaps of time in which the talking heads don’t have much to talk about. In that environment, do you imagine that juicy outliers might get cherry-picked and presented as represented as being representative of the larger electorate? If you say, “Yes,” congratulations: You have watched election night coverage before.

2. As The Upshot’s Nate Cohn notes: “The problems begin early on election evening, when the first waves of exit polls are invariably leaked and invariably show a surprising result somewhere. You’re best off ignoring these early returns.” That’s because they’re not complete, duh! People are still voting, which means the exit pollsters are still polling. ”
3. Even at the end of the evening, exit polling data isn’t perfect. Per Cohn:

The problems continue with the final waves, which analysts pore over in the days after the election and treat as a definitive account of the composition of the electorate. Some foolish journalists might write entire posts that assume that the black share of the electorate was 15 percent in Ohio. In reality, the exit polls just aren’t precise enough to justify making distinctions between an electorate that’s 15 percent black and, say, 13 percent black.

That’s because exit polls are like all other polling! It’s a survey of voters, and like all surveys of voters, they come with the same caveats about sample size and error margins. Plus, they keep compiling the data right up until the end of the night, so if you’re hearing about exit poll results come evening-time, remember that the exit pollsters still have several hours of data left to collect.

4. All the while, there’s the concern that exit polls might affect voter behavior. There’s no concrete evidence of this, but if you’re standing on line at your polling place late in the evening, and you’re already hearing about leaked exit polls that make it sound like one candidate or another has taken a lead, it could cause you to become discouraged. Or, alternately, it could cause you to become overconfident. Either way, maybe getting back to a warm home and a tasty dinner (or a stiff drink) starts to sound more appealing. In 1980, the networks declared that Ronald Reagan had won some three hours before the West Coast polls had closed, and while no one is suggesting that it might has reversed the outcome of the race, concerns that early exit poll data had depressed voter turnout were such that Congress held hearings about it.


A few news organizations that don’t participate in the National Election Pool have gotten creative and curated their own data sources. Some of these will be available all day long as updating models. Slate and Vice News are working with Sasha Issenberg’s VoteCastr project to release continuously updating estimates of where the race stands based on who is turning out to vote so far.

The problem with this information is that it could be really unreliable throughout the day. Nate Cohn detailed the roller coaster ride such a model took the Obama campaign’s analysts on in 2012 ― and there’s a possibility VoteCastr could take everyone along for a similar ride. So you’re much better off to pretend it doesn’t exist if you’re not capable of taking it with the proverbial grain of salt.

And there will be other exit polls besides the official NEP exit polls. These should also be treated cautiously ― even more cautiously, perhaps, since they’re untested newcomers on the scene.


Do not let the data that purports to know who’s winning affect what you do on Election Day. Go to your polling place. Stay in line. Tune out what you think the models and exit polls are revealing. When you get home, don’t look to them as a distant early warning that the candidate you backed is rolling to a win or spiralling to a loss.

Sooner or later, we are going to get the results of ―cliché alert! ― “the only poll that really matters,” also known as “the election results.”

Don’t let exit polls keep you from doing your duty as a voter. And please, please, do not play a role in exit poll dissemination on social media. As best as you can, stay calm, be patient, and make sure you cast your vote.


Natalie Jackson is the senior polling editor of HuffPost Pollster. HuffPost Pollster’s final projection of the presidential race can be found here. Jason Linkins edits “Eat The Press” for The Huffington Post and co-hosts the HuffPost Politics podcast “So, That Happened.” Subscribe here, and listen to the latest episode below.



The United States Votes 2016