THE SAMPLE
Real voters, one question
Voters leaving a sample of polling stations fill in a mock ballot paper. It asks only who they just voted for.
Who is in it?
AN INDEPENDENT EXPLAINER
One question,
asked at the door.
The exit poll asks people who have just voted.
Interviewers stand outside a sample of polling stations and hand voters a mock ballot paper. A team of statisticians and political scientists turns the answers into a seat forecast, published the moment polls close.
10PM not a minute before
Who is counted?- Ask real voters
- Measure the change
- Forecast, not result
- How the exit poll works
THE SHORT ANSWER
Sample, compare,
then model.
The exit poll is not a count.
It is a careful estimate of change since the last election.
THE COMPARISON
Same places, next time
Many polling stations are revisited election after election, so the poll measures change against earlier exit polls at the same spots.
How it is built
THE FORECAST
Chances, added up
A statistical model gives each seat a probability for each party. Adding them up gives the expected seat totals.
How close it getsGeneral information only, not legal advice. For advice on your own situation, speak to a qualified lawyer or a free advice service.
FOUR STEPS ON ELECTION DAY
From the doorstep
to the forecast.
The fieldwork is done by a polling company; the analysis by a small team of academics who work all day in a sealed room, phones handed in, until the 10pm broadcast.
Choose the stations
Reuse stations polled last time where possible, replacing those whose electorate or layout has changed.
Ask one question
Voters leaving the station fill in a mock ballot. Only people who have actually voted are asked.
Measure the change
Compare each station with its own previous exit poll, because UK votes are counted by constituency, not by station.
Model every seat
Apply the estimated change to each constituency's last result, turn it into win chances, and add them up.
TWO NUMBERS WORTH KNOWING
A small sample,
a careful model.
One number on how many places are visited,
one on how close a good forecast should be.
130+
Polling stations visited
Teams stood at more than 130 polling stations across Great Britain at the 2024 general election. Earlier explainers describe about 100 stations out of some 40,000.
FIELDWORK · 2024
≈ 20 seats
Reasonable error on the majority
A well-run exit poll's predicted Commons majority might be expected to land within about 20 seats; within 10 counts as exceptionally accurate.
ACCURACY · UNIVERSITY OF WARWICK
Before the current method, errors were larger: in 1992 the two broadcasters' forecasts of the winning majority were out by 62 and 70 seats. With the newer method the error was 0 in 2005 and 2010.
WHO IS IN THE POLL?
Is this voter
in the exit poll?
Pick a situation to see how the exit poll handles it. The bar shows how directly the poll measures that vote.
General description of the method used at UK general elections since the early 2000s. Not a description of any one election's figures.
THE 10PM RULE
Why you never see it
before polls close.
Under section 66A of the Representation of the People Act 1983, nobody may publish a statement or forecast based on what voters said after voting until the poll has closed. It covers parliamentary elections and local government elections in England and Wales.
Breaking the rule can lead, on summary conviction, to a fine or up to six months in prison. That is why the forecast is released at the close of polling at 10pm, and not a minute before.
Questions
Questions people ask
How is an exit poll different from an opinion poll?
An opinion poll asks people how they intend to vote. An exit poll asks only people who have just voted, so it records behaviour rather than intention.
Does the exit poll say who wins each seat?
No. It gives each party a probability of winning each seat. Adding those chances up across all seats gives the expected number of seats for each party.
Why go back to the same polling stations?
UK votes are counted by constituency, not by polling station, so there is no official result for a single station. Comparing with an earlier exit poll at the same place is the best way to measure change.
Are postal votes included?
No. People who vote by post are not in the exit poll. The method's authors say this has not yet been a major problem but could become one if postal voting keeps growing.
Can an exit poll be published during voting?
Not at UK parliamentary elections or local elections in England and Wales. Section 66A of the Representation of the People Act 1983 forbids publishing it before the poll closes.
Sources
- University of Warwick, Department of Statistics: Exit polling explained
- Ipsos: What is an exit poll and how does it work?
- legislation.gov.uk: Representation of the People Act 1983, section 66A
- UK in a Changing Europe: The exit poll, how it works and why we need it (2024)
Updated