Wednesday, February 18, 2015

Swashbuckling SNP seize 17.8% lead in splendid Survation survey

This month's full-scale Scottish poll from Survation has just been released...

Scottish voting intentions for the May 2015 UK general election (Survation, 12th-17th February) :

SNP 45% (-1)
Labour 28% (+2)
Conservatives 15% (+1)
Liberal Democrats 5% (-2)
Greens 3% (n/c)
UKIP 3% (-1)

The SNP lead is actually much closer to 18% than to 17%, but is reported as 17% due to a quirk in the rounding.  On the unrounded figures, the SNP have 45.3% of the vote, and Labour are on 27.5%.

If you look at this month's poll and last month's in isolation, you'd conclude that there's no reason to suppose that the SNP's colossal lead has contracted even marginally, because the small percentage changes are consistent with margin-of-error noise.  However, we do have to bear in mind that - incredibly - last month's 20% gap was the lowest SNP lead Survation had recorded since the referendum (unless you count the September 19th telephone poll).  So a further small narrowing does open up the possibility that Labour have made some paltry inroads over the last two months - but it certainly doesn't prove that beyond all doubt.  It's still conceivable that there's been a steady SNP lead of roughly 20 points since October, and that the variation between 18 and 24 points that Survation have recorded has been caused by the margin of error.  One reason for thinking that might be the case is that the last poll from the firm saw a drastic upweighting of a very small SNP-heavy sample of under-25s, and that sort of thing can increase the chances of volatility from one poll to the next.

The other important point to make is that the evidence for the SNP's vote having dipped is significantly weaker than the evidence for Labour's vote having recovered slightly.  Survation's first post-referendum poll for the Record put Labour on 24%, so a 4% increase since then might just be statistically significant.  But the SNP were on 46% in the same poll, and the 1% drop since then is utterly meaningless.  Nicola Sturgeon's party have been hovering at 45% or 46% in three of the four polls in the Record series to date, with their 48% showing a couple of months ago looking like the obvious outlier.  So on the face of it, what we're seeing is the SNP's support remaining more or less untouched, while Labour hoover up a few soft votes from the smaller parties - perhaps from people who are impressed by Jim Murphy's Tory Lite credentials.

If there's any truth in that, it means there's a big danger for Labour - because even assuming a small recovery has occurred at all, it may have occurred in the wrong geographical areas.  The potential for "Murphy Tories" or "Murphy Kippers" to turn things around for Labour in Glasgow or North Lanarkshire is obviously pretty limited.

The ticking of the clock was of course the Yes campaign's greatest enemy during much of the referendum campaign, but it's become a bosom friend for the SNP now.  If the recent polls from Survation, YouGov, ICM and Ipsos-Mori have been accurate, then the odd 2% trimming of the lead here and there isn't going to be anything like enough for Labour with only two-and-a-half months to go until polling day.  The unholy trinity of Murphy, McTernan and McDougall are going to have to hope that one of two things prove to be true - either a) the intensity of the official campaign period will offer an opportunity for a game-changer that has so far eluded them, or b) the two Labour-friendly pollsters (Panelbase and TNS-BMRB) are right and all the other firms are wrong.  If Labour were only 10 points behind in January rather than 20, then obviously a 2% narrowing of the gap would be more significant, especially given the way the electoral system works in Labour's favour when the popular vote is relatively tight.  The next Panelbase poll will be fascinating - if it shows a single-figure SNP lead, perhaps Labour can justifiably feel that there is a small glimmer of hope.  But in truth, we all know that the last Panelbase poll used a very dubious methodology, and as long as there's no repeat of that, a very plausible case can be made for thinking the SNP's lead will actually increase markedly.

As usual, the Survation poll also contains voting intention numbers for next year's Scottish Parliament election...

Constituency ballot :

SNP 48% (-2)
Labour 28% (+2)
Conservatives 13% (+1)
Liberal Democrats 5% (-1)
Greens 4% (+1)
UKIP 2% (-1)

Regional list ballot :

SNP 39% (n/c)
Labour 22% (-1)
Greens 13% (+3)
Conservatives 12% (-2)
UKIP 6% (n/c)
Liberal Democrats 6% (-1)

I was about to say that this is the entirely familiar pattern of the SNP being threatened most by defections to the Greens on the list, but in fact there is something unusual here - it's been rare of late to see Labour falling back so sharply on the list.

*  *  *

Political Betting are running a general election prediction competition - amazingly, my entry was accepted, although doubtless it's only a matter of time before I'm disqualified for being too left-wing, or for having too many Irish ancestors, or for the inappropriateness of my eyebrows, or something.  But the interesting thing is that everyone's entry is publicly viewable, and Gerry Hassan (assuming it's the real Gerry Hassan) has made the following prediction -

Labour 291, Conservatives 275, Liberal Democrats 38, SNP 25, UKIP 3, Greens 1

That means he's expecting the SNP to drop back significantly between now and May, and Labour to retain their familiar place as Scotland's largest party in terms of seats (albeit possibly not in terms of votes).  I think he's probably wrong about that, but it does help to put in context some of the bewildering Ian Smart-esque commentary that Gerry has produced since voting Yes in the referendum.

*  *  *

I can't claim to be on the best of terms with the supremo of Better Nation, but as far as I can recall I've never had a cross word with the blog's other regular writer Dom.  To be honest, I'm not entirely sure who he is, but he's undoubtedly an excellent writer.  Just to prove that point, his very funny valedictory post has effortlessly pulled off the supreme feat of insulting virtually every single person in the Scottish political blogosphere, including (indirectly) James Mackenzie himself -

"A few years ago I was invited to become a regular part of Better Nation, at a time when it was the only blog in Scottish politics that did not exist entirely to massage one person’s ego or to cheer-lead religiously for a particular political party...Scotland is a much more complex place than anyone would really care to admit, and what needs to happen now in reflecting that cannot come in the form of a blog, however well intentioned its authors might be. I have never been the kind of person to salivate over polls or write insight pieces just to cultivate my own sense of performative hackery...There are still things to be written, not least over at the Scottish Review and The Conversation, but not here any more."

We love you too, Dom!

59 comments:

  1. No real change, all margin of error stuff. As far as Labour is concerned, it's the difference between getting shot with a single bullet to the head as opposed to a machine gun barrage. They're screwed either way unless something quite miraculous happens.

    ReplyDelete
    Replies
    1. Fantastic way of putting it Stoat.

      Delete
  2. Rounding flatters Labour slightly, because it is 45.3 SNP and 27.6 Labour.

    Slight oddity: SNP are doing a bit better with women (46.1 v 44.6).

    A few yes voters seem to be drifting back to Labour (10% of yes), but this is compensated by the SNP having 11% of no voters.

    ReplyDelete
  3. I'm going to bum myself up and say that this was a damn impressive prediction earlier...
    :)

    ***
    Sweepie for tonight's poll?
    I'll go...
    SNP 43
    LAB 27
    CON 16
    LIB 5
    UKP 3
    GRN 3
    ***

    ReplyDelete
  4. There are Ashcroft constituency polls published tomorrow 11am. He didn't say what type of seats - will be either English marginals or the second batch of Scottish seats that he mentioned in his report a couple of weeks ago.

    ReplyDelete
    Replies
    1. I'm pretty sure they're English seats - he would have specified they were Scottish otherwise (unless he's tacked on a couple of interesting Scottish seats to his English polling).

      Delete
    2. Interesting that Wings has retweeted an SNP supporter (@arcofprosperity) saying that Populus have been surveying East Renfrewshire en masse.

      Populus have done Ashcroft polling, but a follow-up tweet mentions that there were questions about (populist) Labour policies, which makes it sound like an internal (push) poll. Populus were the internal pollster of choice for Better Together. I don't recall any of their surveys re: indyref ever being published.

      re: Ashcroft more specifically, if it is an English marginals batch, he may at least mention how / when more Scottish polling is coming along.

      Delete
    3. It might be an extension of the internal polling that the Guardian mentioned had been going on in Glasgow, and which they vaguely claimed had shown encouraging signs for Labour. (Although if it's push-polling, I'm not sure how reliable those encouraging signs would be.)

      Delete
  5. There is no change to any Scottish polling by any company since the start of November. Nothing beyond MoE at all.

    ReplyDelete
    Replies
    1. This much is very true - every since the 'big switch' the gap in all the polling I've seen has had the gap between LAB and SNP holding remarkably steady.

      Delete
  6. The changes from 2010 to 2015 are quite stark.

    SNP 45% (+25%)
    Labour 28% (-14%)
    Tories 15% (-2%)
    Fib Dems 5% (-14%)

    ReplyDelete
    Replies
    1. I typed Lib Dems but it has appeared as Fib Dem - auto spell check knows best-:)

      Delete
    2. That happened to Dr Eoin Clarke on Twitter too....unless he meant it!

      Delete
  7. Curiously enough. Mike (racist libdem twatmonitor) Smithson claims that this poll shows the trend is all towards labour and the divine Smurph.

    Also Monica of Monza stated that Scotland has an established church. The clue is apparently in the name "Church of Scotland". The whole interweb at his fingertips and still manages to get every single fact he's ever posted on Stormfront Lite completely and utterly wrong. Don't Italian fascists go to school anymore?

    ReplyDelete
    Replies
    1. I may have made this observation before, but Mike Smithson is a buffoon. This poll suggests that Labour MAY have made some extremely limited progress, but you'd have to be utterly delusional to look at it and conclude that Murphy is heading for anything other than a caning in May.

      Delete
    2. At the same time in 2011 the SNP went ahead of Labour and never looked back. I have a feeling the SNP's support is fairly stable and unlike 2010 because of the potential hung parliament will get far more airtime and should help. Plus the growth of social media coinciding with drop in newspaper circulation make it a more interesting election than we would have thought a year ago.

      Delete
    3. In what way is the Church of Scotland not an established church? Parish ministers have to provide pastoral services for everyone living in their parish, not just members. There is a church in every parish, one way or another, with the whole country being covered. It provides chaplains to non-denominational schools and these schools default to Church of Scotland school worship. The Queen becomes a member of the Church of Scotland when she is in Scotland.

      What am I missing here?

      Delete
    4. This is what Wikipedia says -

      "It is the national church of Scotland; however, while it is an established church in the sense that it is recognized as "the Church" in law and its rules, courts and assemblies are part of the law of the land, it is not a state church in the sense of being under the control of the state."

      Delete
    5. Rolfe because it is not established. the CoS has no governmental roll and never has.

      Delete
  8. @Marcia, >"Plus the growth of social media coinciding with drop in newspaper circulation make it a more interesting election than we would have thought a year ago."

    Hopefully, that should make a big difference compared to 2010.
    SNP also in TV debates, plus non-Scottish Labour leader.
    Big changes.

    ReplyDelete
  9. Replies
    1. Yes, but I'll be surprised if there's any Scottish interest this week.

      Delete
    2. Can't someone just leak the results early again so we can find out quick? :P

      Delete
    3. I checked this yesterday. It looks like he has secured the image files directory, which is how the Scottish results were "leaked".

      Delete
  10. "If there's any truth in that [increase in Labour percentage], it means there's a big danger for Labour"
    Polls narrow, new blow for Labour..
    Come on.

    ReplyDelete
    Replies
    1. Yeah, so you've proved that a comment can be rendered less meaningful by changing the words. Congratulations.

      McTernan isn't going to pay you forever unless the quality of your trolling improves, m'laddio.

      Delete
    2. The biggest downward movement in the Survation poll wasn't SNP (down 0.5%), but was in fact the Lib Dems (1.7%). If there is a movement from Lib Dem to Labour then this would logically be more likely to happen where the rump of the Lib Dem vote is (i.e. the Lib Dems have bugger all votes to lose in Glasgow).

      The problem for Labour is that where the Lib Dems (and the Tories) are stronger is normally in places where Labour are not competitive, i.e. the rural areas. There's only a handful of seats in Edinburgh where a Lib Dem (or Tory) movement to Labour would significantly improve Labour's chances of winning seats.

      For Labour to win a larger number of seats, they really need to knock the SNP share of the vote below 40%. FPTP has historically worked against the SNP because they tend to get some votes everywhere, but not enough to win in many places. But with polls at 45%, FPTP would work in favour of the SNP.

      There is a pretty even spread of support for the SNP, even at this higher level of support. The lowest region in this poll is Lothian at about 38%, which is still about 10 points ahead of Labour. The only region where the SNP is well above the 45% average is the north-east, with 56%. All the other regions are in a tight 41%-45% range. The only regions where Labour are above 30% are Glasgow (36%) and West (34%), which is about 10 points behind the SNP in both cases.

      To cut a long story short, Labour need to win over people who voted yes and who now intend to vote SNP. The returns from winning over Tories, the few remaining Lib Dems or SNP no voters is very limited.

      Delete
    3. Indeed, and this problem is compounded by the fact that those traditional Labour supporters that were sophisticated and brave enough to vote YES in September, are not so easily fooled or scared these days. So, two main tactics of the Murphy team (fear and confusion) aren't working so well. They still have a third tactic though - to ensure that their traditional grey vote gets out. Many of those that voted in the referendum won't be so keen to be dragged out again to save the union, but if Murphy and co can pull it off, there are still enough of them (just) to keep their losses down to 10-20 seats IMO.

      Delete
  11. YouGov sub-sample: SNP 43, Lab 26, Con 17, Green 7.

    Almost bang in line (Greens a bit higher this time) with the YouGov averages since November.

    ReplyDelete
  12. Extrapolate, extrapolate, extrapolate!

    Using 3 months of your data James gives a trend that predicts on polling day;

    SNP 44%
    Labour 26%
    Tories 17.5%
    Lib Dems 5%
    Others 7.5%

    The data shows a slight rising trend for the SNP. a slight declining trend for Labour, a slight rising trend trend for the Tories and a slight declining trend for Lib Dems.

    Make of that what you will.

    ReplyDelete
  13. Oh well, the Ashcroft polls were all Tory / UKIP battles. No sign of any new Scottish polls.

    Guess that means the Populus polling in East Renfrewshire is Labour internal shit.

    ReplyDelete
  14. Somewhat off topic but if anyone is interested I've just received the fourth mail shot in three weeks from Mike Crockart (Libdem mp Edinburgh West) after hearing nothing from him for the preceding two years.

    He's clearly worried and equally clearly has plenty of cash backing him. I'd expect an informal pro Union alliance in this constituency.

    Nothing from labour, Tory, snp, or greens yet.

    ReplyDelete
  15. Gerry Hassan is miles away. Out of touch with his recent postings, and not really understanding the polls. He is not a good analyst of polling. His prediction is subjective and lacks scientific reasoning.

    ReplyDelete
  16. luigi -The caveat with getting the grey vote out is thus.

    In the referendum it mattered as there was an 85% turnout. In a GE we will be lucky to get 65-70%. People who turned out to vote no in the referendum will be absent from the GE.

    They were turning up to vote against something not for something. The low turn out suits the SNP and their supporters. They are young,hungry for change and motivated. Labour voters are old,lazy and predictable in their voting turnout which is low.

    ReplyDelete
    Replies
    1. Not sure about this. It was Yes-voting areas that had the lowest turnouts in the referendum, which suggests that being "hungry for change" and "motivated" don't necessarily go together.

      Delete
    2. Point taken although the areas where there were a 'low vote' was areas of population movement, electoral registers very out of date, people registered twice so only voted at one address because they moved from one area to another. This is not a national referendum but 59 electoral contests and providing you candidate gets 1 vote more than any of your opponents in a low turnout area you win.

      Delete
  17. Boab - previous thread. I've been keeping track of James' poll of polls, there's a graph here if you're interested. I'm also tracking what happens if the POP figures get put through electoral calculus (only choosing them because there site is easy to access automatically). In terms of seats, the trend can be seen here.

    ReplyDelete
  18. Yep, I take your point. But remember, Murphy only has to get a few of them out (or voting by post) to reduce the SNP lead enough to hold many seats for the red tories. I expect record numbers of postal votes in May. I agree voter turnout will be lower than the referendum, but it will probably be considerably higher in Scotland than elsewhere in the UK. This is going to be a fascinating GE.

    ReplyDelete
  19. Yes Luigi -The postal votes...ahem.

    These worry me deeply, with SLAB's previous shenanigans on Kirkcaldy!

    ReplyDelete
  20. gerry hassan - the man who wrote a book called the "Strange death of Labour Scotland" - seems increasingly squeamish that this fate is coming to pass. When you see the latest brainfart about how Scotland is missing out on £60 billion of decommissioning work when the oil runs out (during the indyref it was described as a bill not an opportunity) - You realise that Scottish labour has become a rudderless party, governed by a mad spectacle of contradictions, where the false story of yesterday is respun as a missed opportunity the next. All faithfully reported by the msm. This doesn't protect labour as many have said, it increasingly makes them look confused and clueless.

    All this has achieved is to hit Murphy with a reverse honeymoon period with the electorate. He looks increasingly like a lame duck and the man hasn't even been elected yet to represent the party in Holyrood. He is a constitutional oddity. Not allowed to speak for it in Holyrood, under the thumb of the party whip in westminster. Relying on a collection of faceless underachievers to hold the fort. The very definition of an omnishambles. Then they hired a couple of Malcolm Tucker wannabes whose mouths are writing cheques the party can't cash. It would be funny but for the fact that these people really want to have control of our lives.

    It would take a complete reversal of labour to regain the ground the lost. Abandoning austerity, TTIP, putting through serious reforms of the banking sector. Problem is Miliband won't do it and Murphy will have to toe the line as he is still an MP and will be an MP until 2016. Between here and now there is a road full of prat falls and sinkholes all lovingly crafted by Scottish labour themselves.

    ReplyDelete
    Replies
    1. I actually think that Milliband would like to put thru serious reforms of the banking sector but he looks more and more like a prisoner of the Blairites.
      The worst part is that when they get humped, the Blairites will say that it's because they were too leftie when the opposite is the case. Maybe Scotland will save them from the Blairites but I doubt it. They seem too stupid to read the public mood.

      Delete
  21. Interesting that more woman back the SNP than men. Nicola effect? Or JM effect?

    ReplyDelete
  22. I looked back at the polls conducted in Scotland during the six months leading up to the 2010 General Election and found that they overestimated SNP support whilst underestimating labour support. Adjusting for the same degree of error now puts the SNP on 35%, labour on 30%. A far more realistic set of figures.

    ReplyDelete
    Replies
    1. Labour had a 10 point lead by this point (in terms of approach to 2010) which had been growing steadily for nearly 1 year.

      Delete
    2. Support for a Labour majority (as the best outcome for Scotland) was at ~60% in Scotland too. That compares with e.g. 19% in the above survation poll.

      ~50% thought Gordon would make best PM. That compares with 19% for Ed right now.

      If we want to make proper comparisons anyway.

      Delete
    3. And won the election by over 22 points. One would expect the reverse effect this time - with the SNP lead declining in the election itself.

      I think the SNP will still beat labour in terms of votes, just not by the margin that is being predicted.

      Delete
    4. Sorry - didn't see your 2nd reply.

      Delete
    5. I think you're deluding yourself, to be honest. This election is going to be completely different from 2010 because the single biggest handicap faced by the SNP five years ago, namely the leaders' debates, is no longer a problem.

      Delete
    6. The final debate will be a head to head between Cameron and Miliband. And that's the one most people will tune in for.

      There is this idea that 2010 may as well be 2010 BC for all the relevance it has to Scottish politics today. I reject that interpretation. Not much has changed in five years. We had a referendum, the parties assumed the positions we expected they would on that issue, and we got the result we expected. Nothing fundamental has changed. No big important revelations, no serious upheaval in our society. And this is why I expect the result to be closer to 2010 than recent polls would suggest.

      You have posted the results of a poll that says 80% of people want another referendum. Doesn't that ring alarm bells with you that the polls may not be entirely reflective of real life? If this poll were accurate it would suggest the vast majority of no voters want to go through it again. Why? Most I spoke to were like myself - they viewed the referendum as an imposition that brought all of us dangerously close to economic disaster. I therefore give no credence to a poll that says most of them want to do it again.

      Delete
    7. "The final debate will be a head to head between Cameron and Miliband. And that's the one most people will tune in for."

      You're getting ahead of yourself - even assuming the debates actually go ahead, the running order has yet to be announced.

      "I reject that interpretation. Not much has changed in five years."

      Fair enough. If you want to make yourself look unutterably silly, that's your own lookout.

      "You have posted the results of a poll that says 80% of people want another referendum. Doesn't that ring alarm bells with you that the polls may not be entirely reflective of real life?"

      What an extraordinary thing to say. It really is time for you to crawl out of the bunker, my friend.

      Delete
  23. Anon : 3 out of 10 for effort. In many cases, the pollsters' methodologies are radically different from 2010 - Survation are weighting by recalled referendum vote, which reduces the SNP's lead - and yet that lead is still 17.8%.

    ReplyDelete
    Replies
    1. Go on - make it four (that's a pass in some universities...)

      Delete
    2. In the University of Amazon you only get to give a crap product a 1*

      There's many an item you would like to award 0*

      Your contribution deserves the 0* accolade.

      Delete
    3. I guess we'll see wont we? You could well be eating your words in two and a half months!

      Delete
    4. And so might you. Will you come back and eat them here?

      Delete
  24. Another terrific poll. Absolutley no momentum to Labour and from the previous Survation one it seems the 2% gain on the vote is ex Lib Dems (down from 7% to 5%).

    We're still months away, but this is very, very heartning to see. I remember the 19th of September very well and like Billy Kay said on Bateman's last podcast, I think too, it was without doubt the worst day of my life. And I'm very cautious about getting carried away, but this is extremely good for our activists.

    I also heard from another ex Labour (voted Green in 2011 and SSP) who is backing the SNP for WM.

    A few more months to go. Let's keep working hard and no rest til 9.59pm on the 7th of May. Then sit back enjoy a bru or two, or whatever your comfort food/drink is and watch the re-birth of Scotland.

    ReplyDelete