Skip to main contentSkip to navigationSkip to navigation
A counting supervisor opens a ballot box  at The Royal Horticultural Halls in central London
A counting supervisor opens a ballot box at The Royal Horticultural Halls in central London on 23 June. Photograph: Niklas Halle'N/AFP/Getty Images
A counting supervisor opens a ballot box at The Royal Horticultural Halls in central London on 23 June. Photograph: Niklas Halle'N/AFP/Getty Images

How the pollsters got it wrong on the EU referendum

This article is more than 7 years old

It was a bad night for the opinion polls, with few predicting the 52:48 split in favour of leave

It wasn’t just a bad night for Europhiles and David Cameron, but also for pollsters, who misread the mood of the electorate in the run-up to the vote.

Of 168 polls carried out since the EU referendum wording was decided last September, fewer than a third (55 in all) predicted a leave vote.

The actual result on the night came in at 51.9% leave, 48.1% remain. Just 16 of 168 individual polls predicted a 52:48 split in favour of leave.

Polls did give a sense of the swing to leave in the first weeks of June, but edged back to favour remain in the final days before the vote. Just two of six polls released the day before the referendum – those carried out by TNS and Opinium – gave leave the edge.

The graph shows the individual results of 168 online and phone polls carried out between 4 September 2015 to 22 June 2016 omitting don't knows/undecideds

The first and last of the polls that predicted the correct split were taken by Survation, which put the vote at 52% leave, 48% remain on both 4 September and 15 June. The polling company reversed that position in the days that followed; a poll it published on 20 June put the remain vote at 51%.

Polling companies came under intense scrutiny after they failed to call the general election of May 2015. On that occasion, both the pollsters and the bookies failed to predict the Conservative overall majority in the election, although the spread betting markets were closer.

In the run-up to the Scottish independence referendum in 2014, a number of pollsters called a yes vote, although most were correct in saying that Scotland would stay in the UK.

In the run-up to the EU referendum, there were 14 separate occasions between 17 November and 17 June when the polls predicted a dead heat between the sides. (In all cases undecided voters have been excluded in these calculations.)

The What UK Thinks poll of polls, which looked at the average share of the vote for leave and remain across six polls at a time, shows that the pollsters were closest to the result in mid-June, when the polling average put the result at between 51% and 53% for leave.

Bookmakers also got the EU referendum wrong. Odds last week put remain around 1-4, implying an 80% probability of a victory for the pro-EU camp.

Most viewed

Most viewed