As the UK heads to the polls, voters will be anticipating the results, but the exit poll could indicate the winning party ahead of the final votes being counted.

So, what exactly is the exit poll and how reliable is it?

FactCheck takes a look.

What is the exit poll?

The exit poll “interviews voters as they leave the polling station on election day to ask them how they have just voted,” Roger Mortimore, professor in the department of political economy at King’s College London explains to FactCheck. He is also director of political analysis at Ipsos, and a member of the team that conducts the poll.

The UK exit poll consists of just a single question – which candidate they have just voted for.

Its purpose is to predict the number of seats each party will win.
The poll – commissioned by Sky News, BBC and ITV News – is designed by an academic team of political scientists, led by Professor Sir John Curtice of Strathclyde University, and is carried out by the research company Ipsos.

Ipsos goes to about 130 polling stations across the country and talks to around 20,000 people in total. A team of interviewers are outside the entrance of each polling station all day, stopping a sample of voters as they leave and asking them to complete the poll.

If they agree, the voter is given a ballot paper and asked to vote again, the same way as they just did, putting the ballot paper into the exit poll ballot box.

Professor Mortimore said that in past elections, “we have found about 80 per cent agree to take part – a very high response rate for survey research”.

Interviewers choose which voters to approach on a random basis using a fixed interval, for example they might approach every seventh voter, but this varies from polling station to polling station depending on the number of voters expected to vote there.

At regular intervals during the day, the interviewers report back to Ipsos how many votes they have had for each party, and Ipsos then passes on those raw results to the academic team, who build a statistical model from the data to predict how many seats each party will win.

Prof Mortimore said that the “broad principle” is that in each polling station the team is trying to estimate the shift of the votes since the last election. Then, by identifying the differences in behaviour of different kinds of polling stations, “the team can build a picture of what is happening across the whole country, even in constituencies where we haven’t polled”.

What time is the exit poll published?

The projection will be broadcast when the polls close at 10pm on Thursday 4 July.

How reliable is the exit poll and how accurate have previous polls been?

Prof Mortimore told FactCheck that “it is not possible to calculate a formal ‘margin of error’ for the exit poll, because it is not entirely based on random sampling, but the predictions are usually very accurate”.

The exit poll has been run using its current methodology for the last five elections – i.e. since 2005. In three, it called the size of the majority exactly right. In two, the projection was within ten seats of the final result. Its largest “miss” was in 2015, when it projected a majority 30 seats smaller than what ultimately transpired.

Javier Sajuria, reader in comparative politics at Queen Mary University of London, told FactCheck that “as polls go, it is probably one of the most reliable”.

“Even in 2015, when the polling industry suffered a lot of criticism for their failures to anticipate a Conservative majority, the exit poll outperformed everyone else,” he added.

In that election, the seat numbers in the exit poll suggested a hung parliament but David Cameron in fact scraped a thin majority.

However, in 2017, the exit poll “did much better and showed that Theresa May had lost her majority,” Dr Sajuria added.

“In general, I would say that the UK exit polls have been largely successful, and much better than in other countries.”

(Image credit: Vuk Valcic/ZUMA Press Wire/Shutterstock)