EU referendum: The opinion polls are basically 50-50

Our Chief Political Commentator tries to sort out the apparent confusion in recent public opinion surveys

Boris Johnson and David Cameron: the polls are 50-50 between them. Getty

The pound fell on Friday when The Independent published an ORB poll showing Leave 10 points ahead of Remain. Since then two polls were published last night that suggested a closer contest. What is going on?

Opinium in The Observer had Remain ahead by two points, while YouGov in The Sunday Times had Leave ahead by one point, by 50½ per cent to 49½ per cent – it is so close we have to use half-percentage-points.

These confusing results have prompted the usual comments on social media: the polls are “all over the place”; “everyone I know is voting Leave/Remain”; “no one has ever asked me”; “what can you tell from just 2,000 people?”

No, the polls are not all over the place. They show the random variation you would expect from surveys. They are still – for all their flaws exposed in the election last year – the worst way of gauging what people think apart from all the others.

Two ways of making sense of them are to look at averages and trends. The average of the three most recent polls is Remain 49 per cent, Leave 51 per cent.

But these are all online polls, which have been producing findings more favourable to Leave than phone polls, so it is worth comparing that with the average of the three most recent phone polls. That is exactly the opposite, Remain 51 per cent, Leave 49 per cent. (Note: I have added 0.35 points to the Remain figures to take account of Northern Ireland, except for Opinium, which is a UK survey.)

No surprise, therefore, that John Curtice’s poll of polls currently stands at 50-50.

For once, I don’t think it is just a journalistic cop-out to say the polls are on a knife-edge. I think 50-50 is our best estimate of what the polls, taken together, are saying. Which hasn’t changed since 1 June, when I last tried to sum up the state of opinion.

That said, the polls may not be accurate – although we don’t know in which direction, and the pollsters have made repeated adjustments to try to correct for biases. More importantly, opinion – and likelihood to turn out – is likely to change in the next 11 days. I wrote in my column today that I think Remain is more likely to win because late shifts are likely to favour keeping the status quo, and the fundamental question of the economy is in Remain’s favour.

But no one can be sure.

The three most recent polls, all online

YouGov/Sunday Times Remain 49½%, Leave 50½%, fieldwork until 10 June

Opinium/Observer Remain 51%, Leave 49%, fieldwork until 10 June

ORB/Independent Remain 45%, Leave 55%, fieldwork until 9 June

The three most recent phone polls

ORB/Telegraph Remain 51%, Leave 49%, fieldwork until 5 June

ICM/Guardian Remain 48%, Leave 52%, fieldwork until 30 May

Survation/IG Index Remain 54%, Leave 46%, fieldwork until 24 May