While there hasn’t been a lot of voting intention polling in recent weeks, there has been quite a lot of Brexit polling – those organisations campaigning for or against it used the summer holidays to get a good bite of publicity. This included some large polls from YouGov for Hope Not Hate and the People’s Vote campaign showing Remain at 53% and Leave at 47% if there was a referendum now. Today there was a new NatCen poll that showed Remain at 59%, Leave at 41% (though do check the important caveat from John Curtice’s report that the sample itself had too many 2016 Remain voters, so it actually implied a position along the lines of Remain 53/54%, Leave 46/47%) and a Survation poll showing Remain at 50%, Leave at 50%.

In terms of what to make of this, I’d give the same advice on support or opposition to Brexit as I do on voting intention. There are an awful lot of polls asking about support for Brexit, and a lot of people inclined to cherry-pick those which they agree with. Don’t pay too much attention to individual polls (especially not “interesting” outliers), watch the broad trend instead.

There are four regular tracking polls that people should look to to judge whether or not the public have changed their minds (the data is all nicely collected on John Curtice’s WhatUKThinks website here. First there are polls that ask how pople would vote in a referendum now – regularly asked by both BMG Research and Survation using the original referendum question, and using a more generic version by YouGov in their Eurotrack series of polls. BMG have been asking this since late 2016, and where early polls tended to still show more people would still vote Leave, that has gradually changed and since 2017 they have consistently shown more people would now vote to stay. Their EU referendum polls this year have averaged at Remain 49%, Leave 44% (Remain 53%, Leave 47% without don’t knows)

instead.

The Survation series didn’t start until 2017 – since then their polls have varied between neck-and-neck and small leads for Remain. On average this year their referendum polls have shown Remain 48%, Leave 46% (51% Remain, 49% Leave without don’t knows). Unlike the other two referendum polls Survation weight their referendum question by likelihood to vote which, given that previous non-voters tend to split in favour of remain, probably explains the slightly lower remain lead.

The YouGov Eurotrack poll is part of a regular poll across several EU countries on how people would vote in a referendum on their country’s membership of the EU, so doesn’t use the British referendum wording. Nevertheless, the results show a similar pattern to the BMG polling – results late in 2016 continued to show Leave ahead, but since then Remain has been fairly consistently ahead. The average across their five polls in 2018 is Remain 45%, Leave 41% (52% Remain, 48% Leave without don’t knows)

The most regular comparable poll isn’t asked as referendum VI, but is YouGov’s tracker for the Times asking if people think Britain was right or wrong to vote to Leave the EU, normally asked weekly. The pattern should be familiar – in late 2016 the poll consistently showed people thought Britain was right to leave, in early 2017 it began to flip over, and it now consistently finds more people think Britain was wrong to vote to Leave. On average this year 46% of people have said Brexit was the wrong decision, 42% the right decision (without don’t knows, it would be 52% wrong, 48% right).

So while the movement across the polls has not been massive (was are generally talking about a swing of 3 to 5 points from the referendum result), given the closeness of the 2016 result that is enough to mean polls are consistently showing slightly more people opposed to Brexit than in support of it. There is one important caveat to add to this. If you look at the breakdown by 2016 referendum vote you will often find the number of Leave voters switching to Remain is that that much larger than the number of Remain voters switching to Leave (if it is larger at all!), this is because polls generally find those people who did not vote in the 2016 referendum tend to split in favour of Remain.


August is normally a quiet time for polling – partly because the political agenda is often quite bare, partly because both pollsters themselves and the journalists who normally commission public polls will likely be taking their holidays (There’s also a question to be asked about sampling when a fair chunk of the country will be on its holidays, though personally I suspect that doesn’t actually make a difference once it’s spread over a month). It means polling in late August was very light, with only the regular YouGov/Times poll, which was published on Friday and had topline figures of CON 39%(-1), LAB 37%(nc), LDEM 10%(+1), UKIP 5%(-1).

Parliament returns next week, and hopefully a busier political agenda will be equally reflected in some more interesting polling.


-->