So What? The Vote Has Already Taken Place And We Are Coming Out.

[anotheoldgit]

http://www.dailymail.co.uk/news/article-5186797/Britons-Remain-Leave-10-points.html

[Peter Pedant]

The prime minister has made that point to the people who matter - our masters in Brussels

in fact it is clear - The British People have voted to exit and she is the one to negotiate it

whether Parliament MUST have a vote on any treaty is more problematic. Treaty making is clearly a royal prerogative so you can make a sensible argument for NO HoC vote.

Maastricht - MPs put down amendments like - Art 35 'or' shall read 'to or for'
and ..... article 56b - necessary shall be "normally necessary"
and generally tinkering around and making it read better

and Major said - come on boys - this is a whips' issue and is all-or-none
which kinda negates the use of a commons vote

Just a thought on a Sunday morning

[Zacs-Master]

It’s pretty indicative that a lot of leave voters regret not informing themselves of the true facts and procedures which Brexit would entail.
It’s clear it’s not going to be the ‘leave’ which they thought it would be.

[Hopkirk]

If you read the article, there has been little movement in views of those who actually voted. This shift is among those who couldn't be bothered to go to the polling station.

[jackthehat]

Quite.
So the Grumblies, from all sides, will have to content themselves with flapping their gums for the foreseeable...

[emmie]

i voted leave as did many people i know, i haven't changed my mind in the least, we are leaving and that should be the be all and end all of it.

[Khandro]

Fake news! Never trust the polls.

[pixie374]

Rubbish- there hasn't even been a revote. Why would the polls be accurate now, when they weren't before the referendum?

[jim360]

Polls always come with a margin of error, so the central value really is overstated often. In the run-up to the referendum, generally speaking no poll had the margin between the two sides very high. Also, despite the apparent impression to the contrary, rather a few polls showed the Leave campaign ahead -- and, certainly, it seemed that the general trend was for the Leave vote to be increasing as the referendum date approached.

If you account for margin of error, then, the 2016 polls weren't wildly inaccurate. They predicted a close race, probably with Remain ahead, but with a great deal of uncertainty -- which is consistent with the final result.

As for this poll -- well, who knows? I don't think it's unreasonable to see fluctuations like this, especially in the wake of the recent deal in which it looks like we've given rather more away than some people were hoping for. But on its own it doesn't mean anything.

But referendums are a snapshot of public opinion at the time of the vote, and that's all, and there is certainly every possibility that opinion either has changed, or will change, since the referendum proper.

[pixie374]

If the poll was for exactly the same people as previously and asked "have you changed your mind since"... it may be a tiny bit relevant. Otherwise, it means nothing.

[jim360]

Tracking the same people in a poll is something you can do but really isn't necessary to get a sensible polling result. In particular, if for whatever reason your first sample wasn't as representative of the whole population as you thought, then sampling the same people again and again will just lead to the same unrepresentative results. So it's probably more sensible to try and sample randomly each time. If you do your job properly, that will still be reasonably representative and can still track changes.

But the main point is that no poll can be 100% accurate, so you need to have a lot of them in order to get a proper meaning. This one poll indeed doesn't say much at all. If it got repeated several times in the course of the next few months, and by different pollsters, then you should probably start to sit up and take notice.

[pixie374]

That is true, jim- but in this incidence where you are concerned with changes of mind, there is no way of getting a clue without asking the same people as before.

[albaqwerty]

surely all Pollsters rely on those polled to answer honestly?

If they don't, then any result of any poll is suspect.

[jim360]

Gee, I wonder if pollsters have ever considered that...

[pixie374]

I think statistics are very rarely useful with individual humans anyway. Unless you literally ask everyone- you have to presume patterns- and that just isn't likely to be accurate with people.

[jim360]

Why would that phenomenon be restricted to humans, though?

As long as you properly account for errors then there is no need to sample the entire population and get meaningful results.

[albaqwerty]

thanks for not being sarcastic Jim.

Waste of money.

[pixie374]

It's most restricted to humans and animals obviously. They don't behave in patterns.

[jim360]

I'm not sure I agree with that last assertion, either for humans or for animals. Obviously there's a limit to predictability for any individual, but over a large population it's surprising how predictable behaviour can be.

But obviously there's a long way to go before we understand a lot of human (and animal) behaviours.

[pixie374]

But you could ask an opinion from half the world's population- and you are still no closer to guessing the opinion of somebody in the other half. There has to be a lot of presuming going on?

[jim360]

It depends on what sort of pattern you mean, then. I am not asserting that I can predict the behaviour of one individual with absolute certainty under any circumstances (although, given enough information, I could probably do better than 50/50). But if you want to know the behaviour of an entire population, or at least of most of it, then it's possible to make predictions on that scale with a much higher degree of certainty. Over large groups, behaviour is much more predictable than a "no pattern" assumption would lead you to.

It's never going to be perfect, of course. But population sampling done sensibly is much more informative than people give it credit for. The real problem is that no-one in the media ever seems to care about the errors -- and, despite the lessons that central values in polls are meaningless without those errors, this basic mistake persists.