Thursday, October 2, 2014

SNP extend lead to 10.3% in Scot Goes Pop Poll of Polls

I went to bed last night convinced that the SNP's extraordinary run of being in the lead in every single Scottish subsample since referendum day had finally been broken.  The tally for 'others' in the headline YouGov numbers was down by 2%, making it quite likely that the combined SNP/Plaid Cymru share of the GB-wide vote was down from 5% to 3%.  And so it has proved, but incredibly that still leaves the SNP narrowly in the lead in the Scottish subsample, with 33% of the vote to Labour's 30%.

It remains to be seen whether this narrowing of the gap is just an example of the wild fluctuations you can expect when comparing individual subsamples, or whether it's the first sign that the post-referendum surge for the SNP is tailing off just slightly.  In the meantime, the SNP's lead has actually increased in this blog's Poll of Polls, which takes account of all subsamples from the last seven days (four from YouGov, one from ComRes, one from Populus and one from Ashcroft).

Scottish voting intentions for the May 2015 UK general election :

SNP 38.4% (-1.5)
Labour 28.1% (-2.2)
Conservatives 17.1% (+1.5)
Liberal Democrats 7.1% (+0.3)
UKIP 4.9% (+0.5)
Greens 2.7% (+0.3)

(The Poll of Polls uses the Scottish subsamples from all GB-wide polls that have been conducted entirely within the last seven days and for which datasets have been provided, and also all full-scale Scottish polls that have been conducted at least partly within the last seven days.  Full-scale polls are given ten times the weighting of subsamples.)

28 comments:

  1. Could you possibly graph your results? It might make it easier to see underlying trends.

    ReplyDelete
  2. I had a look at the average SNP lead in the previous five YouGov polls yesterday - it was at 17%. Putting it into the seat calculator at ScotlandVotes and it gave the SNP 43 seats, Labour 12, Conservatives 2 and the Lib Dems 2.

    ReplyDelete
  3. Anon : If anyone could patiently talk me through step-by-step the process of setting up a decent-looking graph, I'd be willing to give it a go. I've tried a few times before and always gave up in despair.

    ReplyDelete
  4. Is it not true that the polls typically underestimate the SNP share of the vote, but also when extrapolating this trend from a subsample is fraught with errors
    However it is balanced by all the polls show the same trend ,It will be interesting when a scot only poll is done (soon ? hopefully ?)

    ReplyDelete
  5. Just came across this blog last night and haven't laughed so much since the Clackmannanshire result came in on the 19th.

    Is this the bubble you guys live in? Everyone as delusional as each other, selectively taking pro-independence bits of articles, highlighting via cognitive bias the bits which back up your case whilst ignoring the bits that don’t, then chucking it out there for a few rounds of Chinese whispers to make a puddle appear an ocean, all followed by the most uncritical commenting I’ve seen to follow-up.

    You accuse the mainstream media of bias, and rightly so, but heavens above, this is incredible. Pretty much every post and every comment is completely unquestioning. It just reinforces the perception that you guys are not a credible political cause, but a cult. Heaven’s Gate, the Peoples Temple, Yes Scotland.
    If you all dared to step outside your bubble of mutual back-slapping you guys would realise you have a serious image problem with a large slice of the electorate.

    ReplyDelete
  6. @ylee
    Historically, polls have tended to overestimate the SNP. The only election I can recall where they outperformed the polls was 2011.

    In any case, I'd be interested in whether Westminster subsamples eight months out have ever come anywhere near the eventual result. I kind of doubt it.

    ReplyDelete
  7. Those numbers give us, using http://www.electoralcalculus.co.uk/
    3 Con
    16 Lab
    2 Lib
    38 SNP
    Wouldn't THAT be fun :D

    ReplyDelete
  8. "you guys would realise you have a serious image problem with a large slice of the electorate"

    Is that what you call yourself? How pompous.

    A gentle hint - it might be a good idea to practice what you preach. Even by the standard of the trolls we've been infested with from your side of the argument, your post was a tad over-the-top.

    ReplyDelete
  9. "If anyone could patiently talk me through step-by-step the process of setting up a decent-looking graph, I'd be willing to give it a go. I've tried a few times before and always gave up in despair."

    It depends on how decent-looking you mean but I'd be happy to set up another chart in Google Docs if you like.

    ReplyDelete
  10. In cults, criticism isn't welcome. Neither is perspective nor questioning.

    Your reply completely vindicates this original point of mine.

    Thank you.

    ReplyDelete
  11. @Anon
    A tad unfortunate that my comment questioning James's post came one minute after your comment stating that no one around here ever questions anything.

    ReplyDelete
  12. "In cults, criticism isn't welcome. Neither is perspective nor questioning."

    And all this from a person whose comments haven't been deleted. Tell you what - try posting a pro-independence comment on a Labour website, or on an anti-independence Facebook page, or on the Adam Tomkins blog, and see how long it stays up - in the unlikely event it ever gets published in the first place, that is.

    Then come back here and try to lecture us about "cults" and "bubbles" with a straight face. (Or even, dare I say it, about self-awareness.)

    ReplyDelete
  13. @Anon

    It just reinforces the perception that you guys are not a credible political cause, but a cult. Heaven’s Gate, the Peoples Temple, Yes Scotland.

    Not sure what to make of this bizarre statement. If we were a cult then we would all be dead by now, because of the No vote. What a utterly batshit crazy statement. The No campaign have had decades to come up with a positive case for the Union, and we are still waiting to hear it. Instead of calling independence supporters deranged cultists, why can you not come up with one positive reason for Scotland remaining a dependency? Go on give us positive reasons, which do not entail the usual catastrophes that would somehow entail from Scottish self government.

    ReplyDelete
  14. Hmm, for some reason my comments are appearing under my "oldchap" handle rather than "Sandy Brownlee" like I've asked blogger to do.

    On the trolling front - this is far from a cult! Yes, there is a pro-independence slant, but you try finding a site that isn't biased one way or the other. Conclusions are always backed up by numbers from polls. You are (and others have been in the past) welcome to disagree if you can back up your arguments. That's healthy debate.

    ReplyDelete
  15. I made a graph of this a few days ago. Will update it and post it later. Problem with it is however, that these subsamples go up and down fairly dramatically.

    ReplyDelete
  16. Historically, for yougov subsets, you can get a labour share up to 24 points different from the previous day. SNP 18 points different. Hence expect wild swings from day to day; a symptom of it being subset data.

    SNP had already drawn level in the subsets with Labour prior to referendum after slow steady gains over the past year+ leading up to it.

    ReplyDelete
  17. Oops, in fact 24 point changes for both parties (excel typo), i.e. Labour can get 44 one day then 20 the next.

    ReplyDelete
  18. "If anyone could patiently talk me through step-by-step the process of setting up a decent-looking graph, I'd be willing to give it a go. I've tried a few times before and always gave up in despair."

    Oldchap's suggestion of Google Docs is a good one, or you could use Excel and exponential smoothing-

    http://www.excel-easy.com/examples/exponential-smoothing.html

    Not sure which is better though, as I've not used Google Docs much.

    ReplyDelete
  19. I'd be reluctant to accept that the SNP vote is under represented in these polls after pinning my hopes to the Yes vote being under estimated ahead of the referendum :(

    ReplyDelete
  20. Google docs is mainly useful because it's free and quite easy to share. It doesn't have as many features as Excel.

    Saying that, smoothing should be less necessary with the poll of polls as that already averages out the worst of the variation.

    I've given it a go with the three points we've had so far:

    http://tinyurl.com/okbayxo

    ReplyDelete
  21. Conversation at work today with an older No/DevoMax voter.
    "That c**t Broon jist taen the pish oot o me, huz e no?"
    I was forced to agree. Lots of angry older Labour voters out there.

    ReplyDelete
  22. I wonder what YES would score now in a new indy poll

    ReplyDelete
  23. SNP ahead in Panelbase poll for Westminster and Holyrood:

    http://www.snp.org/media-centre/news/2014/oct/panelbase-poll-snp-westminster-and-holyrood-lead

    Interesting that the Westminster VI was apparently 2010 weighted. Could be underestimating SNP vi. We should be able to see from the tables.

    Also holyrood VI was EU election weighted so may underestimate SNP again.

    ReplyDelete
  24. http://111.imagebam.com/download/7J_FXmXydccMrhawmin4Ew/35514/355139840/Screen%20Shot%202014-10-02%20at%2015.13.16.png

    I made one mistake with the graph, listing the latest SNP poll at 40, when it should be 33 - the rest is right though.

    ReplyDelete
  25. Very interesting to see a full scale SNP lead in the WM elections too.

    Hopefully the SNP can keep up the pressure.

    We need devo max.

    ReplyDelete