Monday 17 October 2011

Mildly interesting, totally irrelevant

From The Metro:

A ComRes poll for the Independent on Sunday and Sunday Mirror found 39 per cent of voters believe Scotland should be a separate country, while 38 per cent disagreed. When the same question was asked in May, only 33 per cent of Britons backed independence for Scotland.

Among Scottish voters, 49 per cent were in favour and 37 per cent disagreed. In May, the figures were 38 per cent for and 46 per cent against independence....


So much to mildly interesting.

... However, a spokesman for the Scottish Office suggested the poll's findings did not reflect the true picture, as only 176 of the 2,004 people surveyed were from Scotland.

The Scottish Office are clearly complete jokers. Scotland's population is just over eight per cent of the total UK population, so if you are interviewing 2,004 people, it seems correct to interview just over 160 people from Scotland (so if anything, they interviewed slightly too many from Scotland).

16 comments:

john b said...

The point's more that if you poll 2000 people, you get a decent statistical confidence level; if you poll 174 people, you don't. So while the poll's likely to be more or less accurate for the UK, the Scotland number will have ridiculous error ranges attached.

John Page said...

And it's the Scots who will make the decision on this.

Tim Worstall said...

What JohnB said.

Damn near regardless of the population size (OK, obviously not if the population size is 500 people) you want to poll 2,000 and no more but also no less in order to get something with statistical significance at the 0.95 level (0.05, depending which way around you want to do it).

Old BE said...

What John Page said.

It is utterly irrelevant what anyone in England thinks. Only the Scots will decide!

Mark Wadsworth said...

JB, TW, these opinion polls are not particularly accurate anyway, to wit the jump from 33/38 in favour to 39/49 in favour in the space of five months, so much to "statistical significance".

Depending on how carefully the sampling was carried out, I think that 174 is quite enough to give an answer to within five per cent either way. That certainly doesn't mean that people will actually vote this way in an actual referendum.

JP, BE, quite true (the biggest stumbling block is whether Westminster will let them and not what English people think), but that doesn't detract from my second point.

Mark Wadsworth said...

JB, TW, to put it another way, I constantly run Fun Online Polls.

For sure, the opinions expressed are not representative of "the UK as a whole", they represent a much narrower category of "people who read this blog who can be bothered to vote", of which there are about a hundred each week.

What I have noticed is that if after the first day's polling, let's say for example, twenty people have voted and let's say half vote 'yes' and a third vote 'no' and a sixth vote 'other'. By the end of the week, and after a further eighty votes, the split tends to be much the same, i.e. half, a third, a sixth (or whatever).

OrneryPest said...

But historically, since it was originally the king of Scotland who inherited the English throne, doesn't that make England, in effect, a territory of Scotland? So it would be up to Scotland to grant independence to England.

Anonymous said...

I'm sure that giving England to chance to leave Scotland to it's Darian Projects would be a winner...

AC1

john b said...

JP/BE: To a point. If England had a large majority in favour of English secession from the Union, that would make a difference.

Mark: I've previously run serious, representative online polls for companies. Agree that it's rare for the split to totally and dramatically change, but going from 45/55 to 55/45 isn't at all unusual.

Ornery: no, in short. The Acts of Union (both the 1707 and 1808 ones) make clear that GB (1707) and then the UK (1808) is a country and a kingdom in its own right, with none of the historical kingship mattering in the least.

Anonymous said...

Irrespective of the views of the Scottish diaspora or other people living in the UK, it will be the Scots and Scots alone taking part in the vote for independence, what is the use of poling those that will have no influence on the outcome.
daft, I call it.

Mark Wadsworth said...

OP, exactly.

AC1, as I've been saying for over a decade.

JP, that sort of shift is "not unusual" i.e. any poll is only accurate to within a ten per cent margin of error. Look at polls held a year after the last GE< about ten per cent of people claim to have changed their mind.

Anon, hence the post headline.

john b said...

Mark: I meant in the same sense as you were talking about for Online Fun Polls - equally for serious polls, as more responses come in over a few days, a 10% swing from first 100 responses to all 2000 is feasible.

Published opinion polls rarely do that unless there are serious methodological differences - to get a 10% Lab-Con or vice versa swing in two polls two days apart would require either an epic screw-up by one polling company, or for Ed Miliband to be caught with Maddie in his basement.

Mark Wadsworth said...

JB: "a 10% swing from first 100 responses to all 2000 is feasible"

Yes.

So the 39% for 2,004 'all of UK' voters is more reliable than the 49% for the 174 Scottish voters; in turn, we'd expect to see support for 'independence' to be higher in Scotland, so the 49% can't be far off.

dearieme said...

"it will be the Scots and Scots alone taking part in the vote for independence": of course it won't. It'll be the voting residents of Scotland, whether they be Scots, Poles, Englishmen, Pakistanis or Celtic supporters.

DBC Reed said...

Not sure that the majority for Scots independence can be expected to be higher in Scotland as suggested above. As a pro-Scot Englishman ,I am surprised to find that a lot of English want to be shot of Scotland prontissimo.
In fact the Scots Nats would do well to include the English in any poll on Scots Independence I'm afraid .

john b said...

Dearieme: Scots and Englishmen living in Scotland as UK citizens, and Pakistanis living in Scotland with permanent leave to remain in the UK as Commonwealth citizens, yes. Not Poles - they can't vote in UK or Scottish Parliament elections, so very unlikely to be included in the referendum electorate.