With three weeks until the result, can we trust any of the EU referendum polls?

The 2010 General Election - The British Public Go To The Polls
Different polls have told very different stories about the likely outcome of the referendum (Source: Getty)

John Curtice, senior research fellow at the NatCen Social Research and The UK in a Changing Europe initiative, says Yes.

In many respects, the message of the polls has been consistent throughout this referendum campaign. Ever since the question that will appear on the ballot paper was settled last September, polls conducted over the internet have on average been unwavering in their insistence that the outcome looks very close. They have consistently pointed to a 50 per cent vote for Remain and a 50 per cent vote for Leave. Meanwhile, since February at least, those polls conducted by phone have also been telling an unchanging story, albeit a somewhat different one. They have on average credited Remain with 55 per cent, Leave with 45 per cent. Of course, this does leave us with the difficulty as to which set of polls to believe – and nobody has yet been able to prove which is right and which wrong. However, that still leaves us with one crucial message that we can and should take away from the polls – David Cameron is facing a much tougher battle than he ever envisaged.

Len Shackleton, professor of economics at the University of Buckingham, says No.

The leading pollsters have defensible methodologies. YouGov, for instance, has a very careful approach to selecting its internet sample. I’m marginally more inclined to their approach than telephone responses, which have a lot of problems now that landline use is declining very rapidly, and not necessarily in a predictable manner. But I’m pretty sceptical of all the polls at the moment. People are becoming jaded with all the garbage they are getting from both sides, and if asked for a spot opinion will produce knee-jerk responses which don’t necessarily reflect settled opinions. Much depends on turnout, and while responses are weighted to reflect what people say about the likelihood of their voting, this is probably the weakest link in the methodology. The weather and random events such as a terrorist attack or major business failure may tip things either way. More plague-of-frogs warnings from Remainers or fantastic promises about cutting immigration from Brexiteers will surely have little impact.

City A.M.'s opinion pages are a place for thought-provoking views and debate. These views are not necessarily shared by City A.M.