Thursday, May 15, 2014

Support for independence at 44% in new Survation poll

I thought we were getting the new Survation poll last night, as it turned out it's tonight!  However as far as I can see all that has been revealed so far are the numbers with Don't Knows excluded -

Should Scotland be an independent country?

Yes 44% (-1)
No 56% (+1)

I've already had to update this post, because the poll was misleadingly billed on Twitter as a no change affair.  The confusion comes about because some people are comparing it to the last Survation poll for the monthly Daily Record series, which did indeed show figures of Yes 44%, No 56%.  However, the correct comparison is with the more recent Survation poll for the Sunday Post, which had Yes one point higher.  It doesn't make a huge amount of difference, though, because the changes between the three polls are well within the margin of error, and offer an impression of recent stability following the winter slump in the No lead - although that's only an impression, and it's not one that all other pollsters agree with.

We'll need the headline numbers that don't exclude Don't Knows before we can really assess where this poll leaves the state of play.  If past form is anything to go by, we may not get those until the morning, and I won't be able to update the Poll of Polls until we do.

Still, the initial reaction of the usual suspects to this poll has been highly instructive.  The glee with which Murdo Fraser claims "more proof of no Yes momentum" betrays a No campaign that is entirely on the defensive, and that has given up all hope of making progress of it own.  That's in spite of the fact that the last two or three weeks should have marked some kind of fightback for No, given the number of different advertising campaigns that have been rolled out (including from those adorable London Tory millionaires in the "Scottish grassroots group" Vote No Borders).  Unlike the TNS poll last night, Survation's fieldwork is likely to be bang up to date, and indicates that the huge spend on anti-independence ads has had no impact, beyond minor changes that are most likely due to normal sampling variation.

And here's a thought for Murdo - if there's such a premium on finding "more" proof of "no momentum for Yes", why don't you simply ask your Tory ministerial colleagues at Westminster to publish the results of the secret Ipsos-Mori mega-poll they recently conducted at huge expense to the taxpayer?  I mean, I know some cynical people say you're only failing to do that because the poll would in fact prove the complete opposite to be true, but surely that can't be right...

The creative geniuses over at McDougall Central have recently developed a fondness for publishing glossy graphs comparing each new poll to the most recent one to be published, even if the firms and methodologies concerned were entirely different.  So we can doubtless look forward in the morning to a graph showing that the Yes vote has increased 3% since the TNS poll just 24 hours ago.  And I gather the Daily Mail has a headline screaming : "Support for separation soars to 44% as outrage over Putin subsides!"

In the comments section below, Rolfe asks if the fact that recent published polls are showing "much of a muchness" means that the secret Ipsos-Mori poll must have been a rogue, and therefore of no great significance.  It's worth remembering that a rogue poll is simply the one poll in twenty that falls outside the standard margin of error, which for a poll with a sample of 10,000 (it may well have been as high as that) would be just 1%.  So in the context of a mega-poll on that scale, even a "rogue" result is actually likely to be fairly accurate - or at least accurate within the parameters of Ipsos-Mori's own methodology and the logic that lies behind it, which may or may not be flawed.  In other words, if the rumours about the poll are true, it would be fairly convincing evidence that the No campaign have lost their most important safety net, and that the only telephone pollster in this campaign is no longer showing them in a near-impregnable lead.  There could hardly be a more significant breakthrough for Yes than that - it would mean that every single BPC-affiliated pollster, regardless of data collection method, is now suggesting that the referendum result is hanging in the balance.

Check back here tomorrow for the headline Survation numbers, and for the latest Poll of Polls update.


  1. Why September 2013? Why does your poll of polls start then?

  2. I think I started it in November. Ivan McKee calculated the numbers back to September and very kindly sent me his figures.

  3. Since these ordinary polls are new and coming out much of a muchness, why is it so important to #publishthepoll? Apart from the fact that we paid for it and they are duplicitous lying swine I mean?

    If Yes is ahead it's just a rogue poll, isn't it?

  4. That would be one hell of a rogue poll, because it would indicate at least a 13% swing to Yes since February! I very much doubt if Yes were actually ahead in that poll - there's probably an element of Chinese whispers in the rumours, with "Yes surge" being turned into "Yes lead". But with Ipsos-Mori being the only telephone pollster in the campaign, it would be hugely significant if they showed a big closing of the gap, particularly in a poll with what seems to have been an unimaginably enormous sample size (which of course reduces the margin of error).

  5. Confirmation that Putingate and the monstering of Salmond by the MSM has hasn't damaged the Yes vote at all?


  7. Stunning new European Parliament voting intention poll
    (+/- change from EP election 2009)

    SNP 40% (+11)
    Con 22% (+5)
    Lab 16% (-5)
    UKIP 8% (+3)
    Grn 7% (n/c)
    LD 5% (-7)
    BNP 1% (-2)

    ComRes - Scottish sub-sample = 172

    That would result in a seat distribution of:

    SNP 3 MEPs (+1)
    Con 2 MEPs (+1)
    Lab 1 MEP (-1)
    LD 0 MEPs (-1)

  8. Mind that Survation asks what you would vote if the referendum is tomorrow rather than in September, over 4 months away. This tends to benefit No slightly although is arguably not as reliable as the referendum is not tomorrow. Asking about September is more likely to be influenced by the way people are leaning even though they may yet have not fully jumped.

    If we get a new panelbase and ICM (ask about September) showing zero change we can more safely conclude things have stabilised over the past few weeks.

    I personally wasn't expecting much change until the impact of the EU referendum registers. That is the next big event and something very tangible. Most people don't follow opinion polls and may not be aware how well UKIP and the Tories combined are likely to do.

  9. I see the record are claiming 'more than 1/3 put off Yes due to Salmond'.

    I suspect a more correct headline is '1/3 of British identifying Scots who don't want independence / prefer status quo don't like Salmond shocker'.

  10. Does that mean 2/3rds aren't put off?

  11. 64% Not put off.

    36% More likely to make them vote No
    12% More likely to make them vote Yes
    46% No influence
    6% DK

    Survation tweet:

    Scottish parliament - constituency vote (ex. ref & undec)
    SNP 44% (- 1%)
    LAB 32% (NC)
    CON 15% (+ 2%)
    LD 5% (- 1%)
    AP 4% (+ 1%)

    Smearing Salmond clearly continues to be a highly effective tactic.

  12. OT, but have you seen this from Maggie Chapman?

    The argument appears to be that the Greens have more than a third of the SNP's vote, and therefore are more likely to grab the sixth seat, so the best way to stop UKIP (my main concern in this election) is to vote Green. I'd be interested in your comment on this.

  13. RE the Greens.

    Aaaaarhg - Yougov sub samples...

    Vote for who you want to represent you in the EU.

    I don't advocate tactical voting and people doing so in large numbers could actually have negative consequences.

  14. In the Yougov sub samples, there is a serious down weighting of the SNP/Plaid votes. Down by half.

  15. Anon : It is highly, highly unlikely that the Greens will be taking the final seat next week. That graph is based on tiny, unweighted subsamples.

    I think there are basically five possible outcomes -

    SNP 3, Labour 2, Tory 1

    Labour 3, SNP 2, Tory 1

    SNP 3, Labour 2, UKIP 1

    Labour 3, SNP 2, UKIP 1

    SNP 2, Labour 2, Tory 1, UKIP 1

    It's therefore possible that voting SNP, Labour or Tory might be the best way of stopping UKIP. There's no way of second-guessing it. But voting Green is only going to help UKIP (barring something very unlikely).