Eve of referendum polls of polls

[eek]

If anyone claims to have a clue what this result will be they are either very brave or lying...

Otto the octupus says its remain however...

[Old Greg]

If anyone claims to have a clue what this result will be they are either very brave or lying...

Otto the octupus says its remain however...

Nigel Farage won't have a clue.

[GB9]

I seem to remember scientists announcing the discovery that the dinosaurs were dying out before the asteroid hit...

Probably the murder of Jo Cox made a few people think about the side they were associating themselves with.

Everybody gets to vote, and only a small minority are loud opinionated people who post on internet forums. We need the moderates to be the voice of reason and as this debate is largely about emotion vs. reason there's still the possibility all the polls are completely wrong.

At this point it feels like waiting for surgery. You hope it'll be okay, but you know at the back of your mind you might wake up with a body part missing.

I think you are correct about the dinosaurs, but there is dying out over thousands of years and there is dying out over a year.

The murder certainly made a lot of people think. If the graphic is correct then it had a massive impact. The graphic shows support for Brexit growing up until the murder and then dropping instantly. There was no sign of weakening until then. The opposite.

The graphic shows support for Brexit strengthening again in the last couple if days but with a lot of undecided. History suggests these will err towards the status quo.

I wouldn't be surprised if the result comes in with remain at 54-46.

[jamesbrown]

If you look on the link attached to the op and scroll down you can see that the average was moving further towards brexit. Then suddenly green (referring to the colouring on the link). The pollsters even reported that voting intentions changed during the poll when the news was announced.

Looking at the op link, you can see the average moving back to red (exit) but not as much momentum as before.

I have no idea how accurate the info on the link is, but at face value, the poor woman's murder made a dramatic change to voting intentions.

You can choose to believe that, but there is no statistical evidence for it that I'm aware of. The headlines in the Mirror and Independent about this swing was fatuous as the sample size in the poll to which you refer was something like 190 after 2PM (when the media began reporting heavily), and there are well-known variations in polling with time of day. The YouGov polling had already started mean reversion on 15-16 June and polled roughly the same numbers immediately afterwards (statistically indistinguishable). I'm not saying you're definitely wrong, but there is no good evidence for it (no direct polling), and any effect must have been marginal to be statistically indistinguishable.

[PurpleGorilla]

My two cents on the potential result.

It will be very close 1-2% difference.

Many leavers have kept very quiet as others have been vilified and called xenophobe. They may come out of the fog and push the leave vote to victory.

Very close, too close to call.

[VectraMan]

When you find out how small the poll samples are the whole thing does seem laughable.

Always makes me think of this:

https://youtu.be/G0ZZJXw4MTA?t=27

[Chuck]

Probably the murder of Jo Cox made a few people think about the side they were associating themselves with.

The mumsnet effect.

They saw the words "Britain First", which the murderer is claimed to have shouted, and immediately associated that with Brexit, not some lunatic nazi group. Rather than look in to it further, they decided that those nasty Brexiters are all lunatic murderers of lovely, beautiful mums.

[GB9]

You can choose to believe that, but there is no statistical evidence for it that I'm aware of. The headlines in the Mirror and Independent about this swing was fatuous as the sample size in the poll to which you refer was something like 190 after 2PM (when the media began reporting heavily), and there are well-known variations in polling with time of day. The YouGov polling had already started mean reversion on 15-16 June and polled roughly the same numbers immediately afterwards (statistically indistinguishable). I'm not saying you're definitely wrong, but there is no good evidence for it (no direct polling), and any effect must have been marginal to be statistically indistinguishable.

Indeed I can't prove I'm correct but looking at the graphic it appears to correlate strongly with the event.

The reversal to the previous trend would further suggest that the one off event was the cause of the sudden deviation.

[jamesbrown]

When you find out how small the poll samples are the whole thing does seem laughable.

Always makes me think of this:

https://youtu.be/G0ZZJXw4MTA?t=27

Oh, it's totally laughable when you see the headlines generated by sub-samples of the headline sample size. Even the headline samples generate large errors, but when you start to talk about changes in the behaviour of particular groups within the sample, it becomes Disney World. You can understand why the papers do it, but even some of the polling companies do this, which surely must be disingenuous (promoting business) if they are statistically trained.

[MrMarkyMark]

My two cents on the potential result.

It will be very close 1-2% difference.

Many leavers have kept very quiet as others have been vilified and called xenophobe. They may come out of the fog and push the leave vote to victory.

Very close, too close to call.

Whereas you haven't kept quiet and you have still being called a xenophobe