Monday, November 5, 2018

Unionist chuntering heard all the way from Chichester after chipper Channel 4 poll gives massive boost to independence

I've been out enjoying Bonfire Night, so I wasn't watching Channel 4's big Brexit show, and I'm trying to make sense of the Scottish figures from their Survation mega-poll, based on the limited information available on Twitter.  Supposedly there are independence figures showing...

Yes 51.4%
No 48.6%

...but I have a sneaking suspicion that'll turn out to be a non-standard question asking respondents to take Brexit into account.  If so, that wouldn't be out of line with similar polls we've seen in recent months, although I certainly wouldn't diminish the significance of it in any way.

I've managed to track down the exact wording of another question in the poll...

From what you have seen and heard so far do you think that Brexit makes it more or less likely that you would vote to support an independent Scotland?

More likely: 38%
Less likely: 25%
Neither more nor less likely: 31%

Which bolsters the impression that Brexit has the potential to secure the small net swing that would be required to produce a Yes majority.

*  *  *

UPDATE: It has been suggested to me by several people that the 51.4% and 48.6% figures are not from a specific poll question at all, but are just extrapolations of what would happen if you adjusted the 2014 referendum result on the assumption that No voters who say Brexit makes them "more likely" to support independence have in fact switched to Yes, and vice versa.  If so, what we're being treated to this evening is the most ludicrous misreporting of a poll that you could ever wish to see.  I can only admire the impudence of whoever came up with the idea.

There are Westminster voting intention figures being quoted from the poll as well, but I think I'll wait to see whether those turn out to be genuine.  On the face of it they show a boost for the SNP.

*  *  *


  1. Interestingly, Curtice claimed that the data from the poll showed Brexit would have no effect either way on support for independence, right at the end of the programme. The data I have seen don't appear to support his words...

    1. Curtice is just a donkey nodding twat who says what the establishment want him to say.

    2. Yes, I thought I heard him say that, and didn't believe it. I think Curtice avoided the direct Indy question and answered one about Remain/Leave that hadn't been asked. Look forward to more detailed analysis of the data re Indy

    3. It's amusing that Curtice is being accused of "avoiding the direct Indy question", considering the way Kerevan and others have misreported this. Why is he wasting everyone's time with this crap?

  2. Has a direct question about independence been asked? As far as I can see people seem to have simply extrapolated the results of the question about being more or less likely to vote for independence. I'd love to be proved wrong.

  3. If you scroll right to the bottom the data tables are available to down load

    1. Maybe I'm missing something but the downloadable spreadsheet seems to just have data on a single question about whether people support a referendum on the brexit deal.

    2. I didn't even know Oral B made a toothpaste!

      Now give me a part in a soap opera.

    3. Download link for ALL tables/Questions

      Sheet tab at bottom - 'Table Index'

      Scotland – Table 6

  4. 'Massive SNP surge, with Labour vote plummeting...'?

    10 hours ago

    Latest @Survation Westminster poll by @Channel4:
    SNP 39.8% (+3.5)
    Tories 26.8% (+0.5)
    Labour 21.5% (-4.5)
    Liberal Democrats 8.2% (+0.5)
    Greens 1.8% (-)
    UKIP 1.2% (-)
    Other 0.6% (-)


    Or rather the last Survation was just at the lower end of variance and the SNP are overall up a tad on 2017, looking to take a few more seats in another Westminster landslide.

    Mind you, if English nationalist Labour stand shoulder to shoulder with the Tories to take Scotland out of the EU come brexit bill day, that will be that for Labour this side of the border. The final nail in the British coffin. Yet England can't back out of brexit now, even if polls say it's increasingly chicken. That would, as noted last night, be the biggest national humiliation in English history. So nope, the great british train crash trundles on, with carriage after carriage derailing in a historic spectacle of utterly screwedness.

  5. I went to a seminar given by air crash investigators. The speaker said one of the main causes of crashes was human error. Usually pilots trying to fix some apparent fault and not watching the whole flight.
    So a red light might be blinking and the whole flight deck can end up concentrating on what was causing it.
    To me we do this with polls. We see the movement of polls like a blinking red/green/red/green light.
    We should be constructing the arguments for independence and working out how to get the overwhelming support for Ind among young age groups translated into votes.

  6. Well, we'll know to take everything Ross Colquhoun says with a pinch of salt from now on.

    1. I haven't taken Survation that seriously since they (and all the other pollsters) concluded ahead of 2014 that weighting to Westminster past vote in Scotland was underestimating Yes so stopped doing this, before deciding to make exactly the same schoolboy error again in recent years.

    2. Oh and Survation UK polling, because it's not Scotland weighted at all (as Survation pointed out to Ross), normally gets SNP much lower than their Scotland-wide polls. The average of the 4 previous UK polls for SNP was 33% for Wesmtinster. This is around 5% lower than they get for full Scottish polls. So, statistical probability suggests the subsample of their latest poll is most likely underestimating SNP by a few %.

    3. This is because UK-wide polls rely entirely on incorrect (for Scotland) westminster past vote weighting while Scottish polls are at least corrected a little by iref/Scottish past vote/EU ref etc.

  7. Regarding the update -

    Looking at the tabs, 17.1% of No voters are more likely to vote Yes and 13.8% of Yes voters are more likely to vote No.

    So the crude extrapolation is:

    55 * .171 = 9.4% N->Y
    45 * .138 = 6.2% Y->N

    Which gives

    N = 55-9.4+6.2 = 51.8
    Y = 45-6.2+9.4 = 48.2

    Not the reported figure.

    1. Interesting to see how you worked that out. Thanks.