Monday, August 18, 2014

Yes vote surges to record high in astonishing poll from the No-friendly pollster YouGov

New YouGov poll :

Should Scotland be an independent country?

Yes 38% (+3)
No 51% (-4)

With Don't Knows excluded, it works out as...

Yes 43% (+4)
No 57% (-4)

And remember these numbers are after the notorious No-friendly "Kellner Correction" was applied.

I'm travelling at the moment, so analysis and a Poll of Polls update will hopefully appear tonight.


  1. Blow for Salmond as nats wrong about Kellner Correction: Boosts Yes.

  2. LOL

    Well, that's awkward.

  3. Assuming the Kellner correction has been applied, what kind of numbers would we have without it? Similar to Panelbase I guess? Or even better?

  4. SS : I haven't had a chance to catch up with Kellner's comments yet - is he claiming the Correction has had a Yes-friendly effect this time?

  5. It's a good poll for yes

    However, as all the posters on this blog have tirelessly strove to point out in the last N weeks: all polls are biased, wrong and prone to error, so we won't be fooled by this one either eh lads? :-)

  6. I took part in this - it was an add on question towards the end

  7. Unionist trolls - I'd like to see a poll with 100% real life participation and weighted to reflect the Scottish Social Attitudes survey...namely 65% of respondents being scottish.

    There are also too many polls with slight changes in methodology for them to be comparable to previous polls.

    It is my belief if the polls were applied correctly, i.e. taking into account how the electorate is actually represented as well accounting for a large amount of 'new voters' we'd see yes level or in the lead by 1-2%.

    Can you tell why the polls done do not reflect the most recent census nor the recent Social Attitudes survey, regarding how people regard themselves?

  8. SS : I haven't had a chance to catch up with Kellner's

    Sorry, I was just making a joke.

    If we get a Yes ahead poll headlines will be 'Blow to Salmond as secret wish for devo max takes poll pounding'.

  9. Hilarious to see this site's now getting trolled 24/7 by BT payroll staffers. Must be doing something right.

  10. Yougov poll, based on fieldwork dates, is 7 point gap closure in 8 days.

    Men now statistical tie, bringing Yougov into line with other pollsters where Men are voting Yes.

    45(+3)% Yes
    47(-4)% No


    Direct swings from No to Yes in each case, from both sexes. Actually larger shift in women, but they were starting further back.

    10/10 Certain to vote is 44(+4)%Y / 56(-4)% N

    Swing is across all age groups, but huge in under 40’s.





    This is in a ridiculous poll which has the SNP behind Labour for Holyrood in a large part due to the ‘Kellner Correction’ which is based on Kellners gut feeling that Scots still support Labour over the SNP. Jeez, even MORI have the SNP comfortably ahead.

    Don’t crack open the champers yet, but if that reflects real movement then things might be just about to get very exciting.

  11. Amazingly biased article based on outright lies in the old labourgraph. From somebody claiming to be and academic too!

    Apparently Yes to change always loses support the closer you get to the vote. He even shows that the Yes vote in 1997 was lower than the support in opinion polls. But we know that's a lie.

    Just another attempt to bolster their failing campaign of hatred against Scotland.

  12. Another thing to consider is that they also weigh based on broad classes of newspaper readership. Given the Sunday Herald is the only one openly pro Yes that'd be another dunt down for the Yes votes.

    I'm still really hazy about how exactly any of these weighings interact.

    Still see the party affiliation question is being reported first. Is it being asked first?

  13. James, Or maybe Scottish Skier (seeing as James is away), can you comment on how weightings applied in this poll compare with other pollsters. It would appear that YouGov consistently have an unweighted panel made up of very old labour supporters. Surely the less representative your panel is the less reliable your overall results are because more weighting is needed. Do other pollsters do any better than Yougov at minimising the need for weighting?

    Also, as a matter of interest surely the different weightings have complex knock on effects across the whole sample - i.e. weighting up under 24's whilst weghting down Labour supporters surely has some cancelling effect. How are the final numbers arrived at, complex iterative solution or something more simple?

  14. Let's hope Alistair Darling can win a few more televised debates!!

  15. I think from the fact that the poll shows labour in the lead for Hollyrood VI YouGov has an overly Labour heavy panel (whether through weighting or otherwise) so would that possibly suggest that because of the excentuation the small swings seen in ICM and panelbase that it is 'traditional' Labour voters we are seeing movement to Yes from? Ie the fact YouGov is overly Labour has excentuated a smaller swing into a large one?

  16. The Labour lead could be another artifact of over correction as far as Labour/SNP switchers are concerned (KC) assuming embarrassed ex-Labour voters aren't admitting to voting for them in 2010. It'd artificially lowering the SNP vote and raising the Labour.

  17. Or it could be that polls aren't a very reliable indicator of how the electorate will vote?

  18. Polls are very good at reporting the views of the folk that are asked.
    Deciding who to ask depends on the pollster.

  19. I'd imagine if there was no such thing as clients with vested interests and pollsters willing to bend their methods to suit that agenda, they'd be much more accurate.

    Sooner or later I reckon when a polling company is coming to the end of a long campaign, they have to get their houses in order and try to get it right, perhaps regardless of what their clients want. The polling companies have to make a living after all this is done too.I suspect they wont want to be singled out as the worst pollsters in what's probably the most interesting political situation in decades!

    Interesting few weeks ahead!

  20. This weighting business by the pollsters reminds me of the trouble I had from my mother-in-law's recipes. I never could get mine to taste like hers.

    When I got to do it with her I discovered that her idea of measuring was to push down the appropriate side of the scale to show the correct reading when she decided by eye that there was enough of the ingredient in the scale pan.

  21. Cynicism might lead one to suspect that as we don't know the fine detail of how weightings are applied it's possible for polling companies to gradually fine tune their results as the actual polling day approaches in order to get as close as possible to the result.

    There's no long term gain in trying to manipulate a poll through results if doing so completely trashes your reputation.

  22. The reputation of a UK polling company is meaningless if you think of it as an arm of the state. The 'UK' company would gladly go burst if it has served its purpose.
    Don't think small. Think of the UK state using every weapon it can. Tanks would be too obvious.

  23. Bit off subject here but this question is coming up time and again on all the pro Indy sites but it's never been answered satisfactorily . How can we be confident that the referendum result can't be rigged ? Can anyone explain the process ? Thanks .

  24. All registered participants are allowed to have monitors at the various stages for ballot box opening, counting etc.

    I would expect our people to be highly trained and extremely vigilant.

  25. Thanks for the reply but I'm just not conviced ? What about the postal votes ? Where do they go ? Who's watching over them ?

  26. It all just seems so vague ! We're going to get robbed here !!!!

  27. Every postal voter must submit a signature for the local electoral register...

    Every postal vote must have an accompanying signature

    As the postal votes are opened and checked a random subset are checked against the register for authenticity and against the signatures

    Someone please correct me if I am wrong...

  28. From the Electoral Commission's website

    WChecking signatures at postal vote openings Introduction This guidance aims to provide practical advice on postal vote opening procedures, with special regard to carrying out manual checking of signatures. It has been developed by forensic experts from the Forensic Science Service, 1along with the Electoral Commission’s own staff, and provides some background to the way electors may write signatures on postal voting statements. This guidance is supported by the Scottish Police Services Authority (SPSA) Forensic Services. It provides three guiding questions that are of use when deciding whether to accept or reject a signature. It also provides a step-by-step decision-making tree to enable people with little knowledge of signature comparison to apply rules consistently. Aim This guidance aims to set out a method for deciding whether to accept a signature on the postal voting statement after comparison with the signature of that elector previously supplied to the Electoral Registration Officer. The guidance will not teach the person looking at the signatures (‘the examiner’) to be an expert, but should help them to make a decision to accept or reject a vote for valid and documentable reasons. In determining the validity of postal voting statements, the Returning Officer or Counting Officer must satisfy himself that the postal voting statement is duly completed and as part of that process must compare the date of birth and the signature on the postal voting statement against the date of birth and the signature contained in the personal identifier record. Although this document contains guiding principles and specific steps that signature examiners should follow when going through the signature checking process, it is for each Returning Officer or Counting Officer to determine whether they can be satisfied that a statement has been duly completed with a genuine signature in each case. Every decision on a postal voting statement should be taken on an individual basis and take into account any other information the Returning Officer or Counting Officer may have. Control signatures 1The Forensic Science Service was a UK Government Company wholly owned by the Home Office with extensive experience in forensic science

  29. If Darling is right about the predicted high level of postal voting made is also as anxious about winning the debate before they're returned as he appears I've a hunch that any attempt at fraud will be exponentially harder to perpetrate given the increased chance that the intended victim might apply by themselves.

  30. I'm unconvinced that postal vote fraud is possible in more than penny numbers. Enough for a council ward perhaps, but for a whole-country referendum?

    Nevertheless the doubt and rumour surrounding the Glenrothes affair is worrying. So much to be concerned about, but all swept under the carpet with neither a transparent explanation nor a re-run of the by-election.

  31. They should have banned postal votes years ago. How many people registered to vote will be genuinely unable to get to a polling place?

    A few thousand across the entire country. Yet more than a million fraudulent postal votes could decide the future of Scotland.

    They must be banned unless you will be out of the country on polling day.

  32. Ed's popular and trusted I see.


    Do you trust ED MILIBAND, leader of the Labour Party, to stand up for Scotland's interests?
    20% Yes
    57% No

    22% Yes
    58% No

    17% Yes
    55% No

    There's that men-women contrast again too eh! Women full of faith in Ed to deliver shiny new powers if we vote No.

    Same level of trust as Cameron. Well done Ed - that's yer currencygate in action!