What are exit polls and how do you read them?

Before the votes are counted, the tallies are released, and the winner is declared on Election Day, the exit poll is the crucial bit of information available to tell you who came out to vote, what candidate they voted for, and why.

Exit polls are conducted as voters leave their polling place, thus the name “exit” polls.

Visit CNN’s Election Center for full coverage of the 2020 race

For the states that hold caucuses, like Iowa and Nevada, exit polls are called entrance polls, simply because interviewers talk to voters as they’re entering their polling place instead of exiting. This occurs because when the caucuses end, unlike a primary, we know the results. It’s more useful to news organizations to have the data to discuss while the caucuses are happening rather than waiting until the end. Because of that, the questionnaires tend to be shorter (a one-sided piece of paper as opposed to double-sided) so that respondents aren’t delayed getting inside to caucus.

Since exit polls are more common than entrance polls, we will focus our explanation on them.

How are exit polls conducted?

Edison Research conducts the poll for the National Exit Pool (a consortium of news organizations who receive their data) to choose precincts and counties in which to conduct exit polls.

As voters are exiting, they’re asked if they would like to participate and, if so, to fill in a questionnaire (a two-sided, 8.5 by 5.5 paper, with 15-20 questions, depending on the state).

Joe Lenski, president of Edison Research, tells CNN that the goal is to get around 100-125 completed interviews with voters at each polling place. They might interview every voter, every other voter or every sixth voter, depending on the precinct.

How and when is the data delivered?

Interviewers call the data into Edison three times throughout the day: after the morning interviews, after the afternoon interviews, and a final call right before poll closing.

What data are news organizations reporting and when?

All news organizations have a quarantine room with a few people inside, receiving data throughout the day with no other outside contact — no phones, no internet, no nothing until 5 p.m. ET. At that time, the quarantine is broken and reporters are allowed to use some of the data as it begins to roll in.

All of the news organizations who participate have an agreement with Congress that they’ve pledged not to report any exit poll results that would characterize the outcome of a race before all polls in that state have closed.

So, for example, in New Hampshire, the last polls close at 8 p.m. ET. At that time, reporters can discuss who is leading in the Democratic primary. Before that, Lenski tells CNN, “You might see questions like ‘what was the most important issue for voters today?’ or ‘what percent of voters were first time voters?’ or questions like that. But any question that would give away the outcome of the race won’t be reported until after all the polls in that state have closed.”

What’re the benefits of an exit poll?

Nationally, Lenski reports that response rates for exit polls are around 40-45%. Compared to a typical telephone poll (response rate around 6-9%) and an online survey (response rate around 1-5%), it’s a benefit to have that higher response rate.

There’s also no worrying about likely voter models. A likely voter model in a survey attempts to define the population that is likely to vote in the election, as opposed to just the overall population. Everyone who participates in the exit poll is a likely voter because they just voted.

What about people who vote on a day other than election day?

Edison has two methods of reaching people who voted early or by mail. They conduct a regular telephone survey in the week or two leading up to the election to reach those who have already voted by mail, specifically geared toward states where bigger groups of the population vote by mail (like Arizona, Washington, Colorado and more).

They also place interviewers at early voting locations in states where majorities vote before election day (like Tennessee, North Carolina and Texas).

What else should a good reader of an exit poll know?

Lenski reminds readers of the exit and entrance polls that the numbers will change slightly throughout the night.

“That’s just because we get more data through the day and as it’s weighted, it gets more precise and refined. We’re just taking the best information we have and refining the results throughout the evening.” After polls close, data is weighted to the official final numbers.

The other question he gets a lot is: “Do people tell the truth in an exit poll?”

He says, “All the information I have is that people who fill out the exit polls are honest with us. The people that don’t want to participate will just choose not to participate. I’ve done this for 32 years and there’s little evidence. People aren’t going to waste their time to lie to us.”