Wednesday, April 8, 2015

Je suis désolé, Monsieur Jackanory, but the SNP's showing in the first post-Easter YouGov subsample suggests the "Frenchgate" smear may not have had the desired effect

After a wait of several days, we at last have the first Scottish subsample from a GB-wide poll to have been entirely conducted after the Telegraph's smearing of Nicola Sturgeon (the previous two were only partly conducted after the crucial moment on Friday night).  The new YouGov poll is bang up-to-date, with fieldwork conducted yesterday and the day before, and the Scottish subsample shows :

SNP 43%, Labour 33%, Conservatives 12%, UKIP 6%, Liberal Democrats 4%, Greens 2%

Labour's share is admittedly at the extreme upper end of their normal range, but there's no reason to suppose that they've taken any votes back from the SNP, who at 43% are polling normally and strongly.  Unfortunately, though, it's much more difficult than before to interpret the meaning of 'normal', because YouGov have quietly revised their methodology for GB-wide polls, and the SNP are no longer being weighted down - which always used to happen as a matter of routine.  (The political weighting is now based on voting intention from a few weeks ago, rather than on antiquated party ID target figures derived from 2010.  Likelihood to vote is also being taken into account, which modestly boosts the SNP from 41% to 43%.)

So we're going to need a few more subsamples to get a clearer picture - but, as of yet, there's still no cause for alarm.

27 comments:

  1. James, do you know when the next full-scale Scottish poll will be conducted?

    ReplyDelete
    Replies
    1. Sam Coates (Times political journo) said last night that there will be a full Scottish poll (by YouGov) in the Times on Friday, so numbers available circa 10pm tomorrow night.

      The methodology change James mentions above (likelihood to vote) should actually help the SNP by circa 1-2 points.

      Delete
  2. SNP were likely still weighted down as I posted on the last thread. Or at least the whole Scotland sample was weighted up and the SNP didn't benefit really from that which is equivalent to a down-weight.

    ReplyDelete
  3. Could a higher labour vote, even if the SNP vote stays the same, not risk a lot of SNP wins? Labour don't need to beat the SNP in the popular vote to hold onto all their seats, right?

    ReplyDelete
    Replies
    1. A higher Labour vote would really need to come from Con. It could help Labour hold some seats, but still lose massively.

      To be honest, a tactical vote from Tory voters for Labour would be disastrous for the unionists.

      If the Tory vote hold firm, even increase a bit, it would allow unionists to argue Scotland was slowly swinging back towards the Tories / union.

      If Tory voters vote Labour, they reduce Labour's losses slightly, but the pattern becomes all unionist parties collapsing in Scotland. That is great for the SNP - voters rejecting all unionist parties increasingly. Labour crashing, Libs collapsing, Tories down... SNP makes big gains is better than Labour crashing, Libs collapsing, Tories recovering...

      It's why Ruth is begging Tory voters to stick with the Tories. She knows how bad it would be for them if their vote falls again at the same time as Labour and the Libs does.

      SNP need 30 seats = majority. Any more would be just a bonus.

      Delete
    2. I should add the above would be particularly bad if Con+Lib+/- e.g. DUP could command a majority. Imagine that happened but in Scotland Libs got e.g. 4(-15)% and Con maybe 12(-5)%. Total democratic mandate for Scotland destruction. Just 16% combined or similar, down from 36% in 2010.

      Nope, Tory voters would be very wise if they stuck with con if they value the union.

      Delete
  4. Anything over 30% for Labour AND under 40% for SNP will enable Labour to avoid a tanking - but we will have to see - don't subsamples fluctuate far more than full scale polls?

    ReplyDelete
    Replies
    1. They fluctuate wildly.

      Delete
  5. The constituancy polling is what is going to kill Labour in Scotland. The national polling is bad for them. But it's when you go into the micro yes voting constituancies. You really see the swings to the SNP. A 16% lead nationally can easily be an 18% lead locally. In some Yes areas it's over 20%.

    I expect Labour to be wiped out in any yes voting areas. I include Coatbridge which they think is a safe seat.

    ReplyDelete
  6. BBC more poll.

    A new iref within the next 5 years?
    48% High priority
    10% Medium priority
    =58% priority
    --
    42% Low priority

    ReplyDelete
    Replies
    1. MORI poll that should be.

      Also, 57% rock solid for devo max. 72% if you include the medium priority people.

      Delete
    2. Although interesting to see that there was actually less support for having a second indy referendum in the next five years than there was for having a referendum on EU membership...

      Delete
    3. Stands to reason. Opposing a second referendum has (almost) become a proxy for opposing independence, whereas there hasn't been a European referendum for >40 years. i.e. there will be a lot of people who intend to vote to stay in the EU who would support having a referendum.

      Delete
  7. I do not understand why this is not cause for alarm, though I recognise it is a subsample and so perhaps not so significant

    If they no longer downweight the SNP vote it should go up, should it not? It hasn't. What would be the expected increase from this? You also say that including intention to vote has boosted SNP in this sub poll. So for comparability should we not deduct both those figures and see what it looks like?

    Perhaps the smear has had an effect? Disappointing if so

    Hopefully I have just misunderstood this

    ReplyDelete
  8. I do not understand why this is not cause for alarm, though I recognise it is a subsample and so perhaps not so significant

    If they no longer downweight the SNP vote it should go up, should it not? It hasn't. What would be the expected increase from this? You also say that including intention to vote has boosted SNP in this sub poll. So for comparability should we not deduct both those figures and see what it looks like?

    Perhaps the smear has had an effect? Disappointing if so

    Hopefully I have just misunderstood this

    ReplyDelete
    Replies
    1. SNP do seem to be downweighted as I posted.

      Highest ever level of SNP identifiers in the unweighted base (big spike), mirroring the same in populus the other day, is no cause for concern. Rather, possible optimism...

      Delete
    2. SS : I can't see where the number of SNP identifiers in the unweighted sample is listed? They've replaced that with the number of respondents who intended to vote SNP in January/February of this year.

      Delete
    3. Ok, I see.

      Same applies then. Number of SNP planners (instead of identifiers) not changed (well, buy one maybe), while the Scottish sample is weighted up as a whole. Ergo, SNP apparently down-weighted.

      Delete
  9. This comment has been removed by the author.

    ReplyDelete
  10. This comment has been removed by the author.

    ReplyDelete
  11. Meanwhile, here's something I threw together on YouTube after watching Jim's performance last night. Maybe some of you will enjoy it.

    https://www.youtube.com/watch?v=7G06aNcy6sI

    ReplyDelete
  12. Just took my first YouGov Scottish political poll in, literally, years. Everything else has been about brands and the like. Questions were voting intention, which leaders are doing a good job, referendum vote (actual and next time), which make-up of government you'd like to see after May and one about whether Nicola did or didn't probably say she'd prefer to see the Tories in.

    ReplyDelete
    Replies
    1. that will be for the one published in the Times on Thursday night / Friday morning.

      Delete
  13. "The political weighting is now based on voting intention from a few weeks ago"

    I assume that is 'voting intention from a few weeks ago' as measured in YouGov polls, right? The same ones in which the SNP were being downweighted, right? So isn't there a kind of downweighting-by-proxy happening, in that the base figures being used were obtained by significant downweighting?

    I could just be talking out of my arse, please do educate me.

    ReplyDelete
  14. Unbelievable James, unbelievable that you're actually trying to spin this as a positive poll for the SNP. This poll/subsample is actually fantastic news for Scottish Labour.

    The closer we get to the election, the more people are warming up to Jim Murphy's leadership. People are forgetting about the referendum and remembering that this is general election and that this is about the United Kingdom. Thanks to the SNP voting with the Tories against Callaghan in 1979, we ended up with Margaret Thatcher destroying Scotland. Canny scots have plenty of memories from that and know that only a Labour majority can stop the Tories for sure.

    If we look at the actual numbers, the SNP used to poll between 45%-50% not that long ago (and that was with them being downweighted). You say that in this poll they've actually been upweighted. This makes it clear to me that voters are abandoning the SNP and going back home to Scottish Labour.

    On top of this, it is useful to remember that on a lot of the Ashcroft polls in Labour safe seats, Labour were only just behind the SNP. If the gap has effectively narrowed from nearly 20 points to 10 points, the swing back to Labour will be enough for Scottish Labour to hold all seats in Glasgow and in other Yes voting areas. There will therefore be no 'Portillo' moments in this election. Jim Murphy, Willie Bain, Margaret Curran, Douglas Alexander, Tom Harris, Ian Davidson and Anas Sarwar would hold their seats and continue in parliament, standing up for Scotland.

    If the trend continues, Scottish Labour can still win on election day. On top of this, I believe that abuse from cybernats and yes voters is making people wary about declaring their support towards Labour. The independence referendum polls had a 'Shy No' factor which I believe will play out in the general election. I believe that there is a 'Shy Labour' factor in place which will probably add 3%-4% on election day.

    Always happy to help.

    P Charles

    ReplyDelete
    Replies
    1. Oh, don't stop there, Mr Charles. Surely you want to mention the words "Tony Blair" and "45 minutes"?!

      Delete
    2. On these figures Labour should expect to lose 50% of their seats in Scotland.

      If this is fantastic news they have surely lost the plot,

      Delete