EU attitudes - a change in political weather?

Earlier today the Times published quite a startling poll by YouGov. Ever since the referendum they've kept a tracking poll ticking over in which they ask the question "Was Britain right or wrong to vote for Brexit?"

Until now there's been precious little evidence for anyone changing their mind. The question you ask matters, of course: if you ask instead something like "Should Britain leave the EU?" then many people who think Brexit is a bad idea will nonetheless tell you we should, given the referendum result.

Other Remain voters make the argument that it's possible for people to change their minds, which is why YouGov have made efforts to track this underlying attitude. And, as I say, up until now there's been very little evidence of a mass change ... although naturally there has been plentiful abuse of statistics to try and argue that there has been in one direction or another. This has more to do with wishful thinking than rigorous statistics.

And so in the past when I've been asked about this polling of attitudes, I've put together a chart of the results with the sceptical position marked for reference - namely, that attitudes have not changed one bit and all the results just reflect random scatter around this fixed position. Here it is with the new result:

The long horizontal lines show the average of all the results over the last year and a half (blue - percentage of people who think leaving is a good idea, red - percentage of people who think it's a bad idea). The vertical error bars around each point have a very specific meaning: 2/3rds of the time, the true result will be within that range, and 19 times out of 20 it will be within twice that range. More intuitively: every poll will be out by a little because you could have bad luck and, say, randomly pick 2000 members of UKIP to poll. But that's unlikely, so polls are usually quite close to the true level of support for Remain/Leave, and occasionally a little bit further away.

In fact, while you can catch maybe a little hint of Remain support dropping by a percentage point or so 200 days after the referendum, all the error bars overlap, or else nearly overlap the average result - which is why I said underlying attitudes had not changed. The exception is this latest poll, in which there's a clear and significant jump in the percentage of people who think leaving is a bad idea.

---

So, how excited should Remainers get? After all, if there's now real evidence for a change in attitudes to Brexit, that will help them pile on the pressure. Well, the truth is that while this result is significant, it may well be a combination of a real movement with an unusually anti-Brexit sample, so it's best not to overstate it, yet. To be certain, more polls will be needed to nail down what's actually going on. That said - and this is a bit more esoteric - attitude changes are not completely random. There's such a thing as political momentum (just ask Jeremy Corbyn...) and there's nothing so persuasive as someone who once agreed with you changing their mind. So these things can sometimes be contagious - but only time (and more polls) will tell.