Exit poll

An exit poll is a survey administered to voters as they leave a polling place on election day, asking who they voted for along with demographic and attitudinal questions. It is used both to help project outcomes ahead of the official count and to analyze how different groups of voters actually voted.

Interviewers at a sample of precincts, chosen to represent the broader electorate, intercept voters and ask them to complete a short questionnaire, often supplemented with phone surveys covering early and absentee voters who cannot be reached at a physical polling place. As real vote counts come in, exit poll results are adjusted through survey weighting to match known precinct-level results and demographics. This differs from a pre-election likely voter model, which forecasts an electorate that has not voted yet; an exit poll instead surveys people after the act.

Exit polls are the main source of the demographic breakdowns reported on election night, how different age, gender, or racial groups voted, and have historically been used, cautiously, to help call close races. The main pitfall is that exit polls have diverged from final certified results in specific elections, partly because voters willing to stop and answer are not a perfectly representative sample of all voters, a form of nonresponse bias, and partly because early and absentee voters, captured through a different response rate and mode, can be harder to survey accurately.