Election Day for the 2026 midterms is Tuesday, November 3, 2026, and within minutes of the first polls closing tonight, TV anchors will start saying things like "according to exit polls, voters' top issue was the economy." But who are these pollsters, who did they talk to, and how can anyone know how millions of people voted before the votes are even counted?
This guide explains how exit polls actually work, how they changed after so many Americans started voting early and by mail, why they sometimes get it wrong, and how to read them like a pro on election night.
What Is an Exit Poll?
An exit poll is a survey of people who have just voted. In the classic version, trained interviewers stand outside a sample of polling places and ask voters, as they leave, to fill out a short, anonymous questionnaire. It asks who they voted for and a set of questions about who they are and why they voted the way they did: age, gender, race, education, income, religion, which issues mattered most, and how they feel about the economy and the country's leaders.
The key difference from a normal opinion poll is timing. A pre-election poll asks people how they plan to vote. An exit poll asks people who already voted, so there is no guessing about who will actually turn out.
Who Runs Exit Polls in 2026?
For decades, the big US television networks pooled their money to pay for one shared exit poll, run by Edison Research for what was called the National Election Pool. In 2025, research firm SSRS acquired Edison Research, and for the 2026 midterms ABC, CBS, CNN and NBC rely on SSRS for their voter survey data, according to Reuters.
The Associated Press, whose vote count is used by newsrooms across the country to call races, runs its own large survey of the electorate as well. Since 2018 the AP has used a different method from the traditional doorstep exit poll: a big survey of registered voters conducted online and by phone in the days before and on Election Day, designed to capture people who vote early, by mail and in person.
So on election night you may see slightly different numbers from different outlets. That is normal: they come from different surveys with different methods.
How Exit Polls Pick Who to Ask
Pollsters cannot stand outside all of the roughly 100,000 polling places in the US, so they use statistics.
- Choose the precincts. Pollsters pick a sample of precincts in each state that, taken together, look like the whole state. They use past results so that the sample includes the right mix of urban, suburban and rural areas and places that lean toward each party.
- Choose the voters. At each chosen polling place, an interviewer approaches voters at a fixed interval, for example every third or every fifth person who walks out, so that the interviewer is not choosing "friendly looking" people.
- Keep it anonymous. Voters fill out the questionnaire privately and drop it in a box. Names are not collected.
- Report several times a day. Interviewers phone or upload results at set times during the day, so analysts can watch the numbers build.
How Exit Polls Handle Early and Mail Voting
This is the biggest change in exit polling in the last 20 years. In recent elections, a huge share of Americans voted before Election Day. The AP's tracker counted more than 57 million advance ballots already returned in the 2026 general election as of early November. A poll taken only outside Election Day polling places would miss all of those voters.
To fix that, pollsters now combine several methods:
- In-person interviews at early voting sites in the weeks before Election Day.
- Phone, text and online surveys of people who voted by mail or early.
- Election Day interviews outside polling places.
The results are then blended and weighted so that each group counts in proportion to how many people voted that way. That is harder than it sounds, because it is not known in advance exactly how many people will vote by mail versus in person.
Weighting: Why the Numbers Change During the Night
The exit poll numbers you see at 7 p.m. are not the final ones. As real vote counts come in, analysts adjust, or "weight," the exit poll so its totals match the actual results. If the exit poll said a candidate was winning a state by 6 points but the counted votes show a 2-point race, the poll is weighted down to fit reality. Every demographic breakdown inside it shifts too.
That is why an exit poll published the morning after the election can look different from the first numbers shown on TV. The final version is the one researchers and historians use.
Can Exit Polls Predict the Winner?
Exit polls are one of several tools that news organizations use to project winners. Race-calling teams combine exit poll data with actual vote counts, past results in the same precincts and their own models. In a lopsided race, a projection can come right after the polls close. In a close race, outlets wait, sometimes for days, until enough real votes are counted.
News organizations also have a long-standing rule not to project a winner in a state until its polls have closed, so that calls do not influence people still waiting to vote. If you want to know when polls close where you live, see our guide to what time polls close on Election Day.
Why Exit Polls Sometimes Get It Wrong
Exit polls are surveys, so they have a margin of error, usually a few points for a statewide estimate and larger for small groups like young voters in one state. A few classic problems can also skew them:
- Who agrees to answer. If supporters of one candidate are more willing to stop and fill out a questionnaire, the poll leans their way. Pollsters call this non-response bias.
- Early vote guesswork. If the mix of early, mail and Election Day voters is different from what the pollster expected, the weighting is off until the real count fixes it.
- Close races. When the gap between candidates is smaller than the margin of error, an exit poll simply cannot tell who won.
The most famous miss came on election night in 2000, when early exit poll data contributed to networks calling Florida for Al Gore, then taking it back, then calling it for George W. Bush, then taking that back too. The networks rebuilt their systems afterward.
What Exit Polls Are Actually Good For
Picking winners is not the main job. The real value of exit polls is the why: which groups voted for which party, which issues drove people to the polls, and how first-time voters behaved. Political scientists use them for years afterward to understand elections. When you hear "voters who said inflation was their top issue broke heavily for one party," that comes from an exit poll.
How to Read Exit Polls on Election Night
- Treat early numbers as a sketch. They are unweighted to the real count, so they can move by several points.
- Look at the sample size. A finding about a small group is much less reliable.
- Compare with the actual count. Real votes always win in the end.
- Remember counting takes time. Big shifts during the night are often just mail ballots being counted later. Our explainer on how votes are counted shows why.
Frequently Asked Questions
Are exit polls accurate?
They are usually close for big groups and lopsided races, but they carry a margin of error of a few points. They are adjusted to match real results after votes are counted, so the final version is more accurate than the first numbers shown on TV.
Do I have to answer an exit poll?
No. Exit polls are completely voluntary and anonymous. You can politely say no, and you cannot be asked to show your ballot.
Can exit polls be released before polls close?
News organizations hold back state projections until polls close in that state. Some general findings about the national mood may be reported earlier in the evening, but not who is winning a particular state.
Why are exit polls done if votes get counted anyway?
Because vote counts tell you who won, but not who voted for whom or why. Exit polls fill in that picture.
When are midterm elections held?
Midterms are held every four years, halfway through a president's term. Learn more in our guide to what midterm elections are and why Americans vote on a Tuesday.
The Bottom Line
An exit poll is a fast, careful survey of people who have already voted, now combined with surveys of early and mail voters. It is a useful first look at election night, and a great tool for understanding why people voted the way they did. But it is a sketch, not the final picture: the counted votes always have the last word.
Comments
Post a Comment