And There You Have It!

[jeffa]

Peter Kellner, head of YouGov, has just given the result, of which he is, in his own words, 99% certain.

And that result is: 46& Yes, 54% No.

And how has he arrived at this? His organisation has polled 1,828 people who had voted, plus another 800 postal voters. That's 2,628 voters out of around an expected 4 million.

http://www.ibtimes.co.uk/scotland-independence-result-no-win-54-votes-yougov-1466175

[Everhelpful]

5 bellies and he is actually English from Newcastle.

[bibblebub]

perhaps a bit of education about statistics and opinion polls would help to understand why only 2,628 voters out of 4 million is sufficient

[Peter Pedant]

// 5 bellies and he is actually English from Newcastle.//

thank you it was an artistic allusion

[is he really from N/c ?]

[puzzlemaniac]

pp - wae aye man

[Peter Pedant]

perhaps a bit of education about statistics and opinion polls would help to understand why only 2,628 voters out of 4 million is sufficien

It is obvious in detecting a 50:50 decision that if the voting were 90:10 either way that your sample would be much smaller to detect the result than if the voting were 48:52. Here they are saying the wobble ( technically related t the standard error of the mean ) is plus or minus 3 on a sample of 1000.

So on a sample of 1000 it will detect anything larger than 47:53
and here we have 46:54.
Here the sample is nearer 3000, so it is gonna detect hlarget than say 48:52....


and why does a sample reflect the make up of the population ?
One of the marvels of statistics
and it will do so even if the population is not bell-shaped ( normally distributed in geek speak)

[Calibax]

@jeffa //If No win, a majority of the Scottish population do not believe that they can run their own country. //
I don't think that's true at all. I'm sure all Scots think that they could run their own country. The 'no' voters just think they would fare better remaining as part of a union which has been immensely successful for 300 years.

[jeffa]

What utter garbage! If all Scots thought they could run their own country better than Westminster could run it for them, they'd have all voted Yes!

Cap in hand does not even begin to describe it!

[mikey4444]

Jeffa...It now appears that Mr Kellner was pretty much right doesn't it !

The percentages are as follows....YES 45%...NO 55%. I have been using YouGovs Polls on AB for a long time now, and this result proves that they have been broadly right so far. Modern Polling is highly accurate these days, and people just need to get their heads around that. By continually throwing cold water on YouGov and others is simply ignoring the evidence.

[jim360]

It certainly is, after all, a triumph for polling. That said, even the pollsters themselves were more cautious than usual over this one, and with good reason. Within the usual errors, though, the polls throughout the campaign now look broadly accurate. A solid base of "no" voters, always there, but when the campaign gets louder people start to turn to "yes" before, at the last moment, turning back after all to the safer path.

[mikey4444]

Lets hope this result for the Pollsters makes people think twice in the future before they dismiss them so readily Jim ! Another thing I noticed in this campaign was the essential shouty nature of the YES camp. It seems that the closer we got to the voting day, the louder and nastier the YES people got.

[-Talbot-]

I seriously doubt that the poll that had the 'yes' vote ahead last week was correct, mikey.

[mikey4444]

YouGov have been pretty consistent Talbot, which was my point. Indeed, remarkably accurate, given the emotive nature of the Referendum.

[-Talbot-]

You can't possibly know how accurate a poll is during the build up.

[mikey4444]

Its this lack of faith in modern Polling that still puzzles me. YouGov have been shown to be 99% accurate...why can't you accept that Talbot ?

[-Talbot-]

I find it hard to believe there was a 23/24% swing in 12 days.

[mikey4444]

There wasn't Talbot...the "don't knows" had control in the end. But most of the Pollsters got it right in the end ! Congratulations to Mr Kellner !

[-Talbot-]

Congratulations anneasquith!

[-Talbot-]

Yes’ campaign lead at 2 in Scottish Referendum

Sat September 6, 2014 3:30 p.m. BST



YouGov referendum prediction: YES 46%, NO 54%

Thu September 18, 2014 10:18 p.m. BST



That suggests a 24%+ swing to me. (I don't think that happened)