Mega Murphy meltdown as SNP rocket to 24-point lead in scorching Survation survey

There's a question I'd rather like one of the nation's journalists to ask Jim Murphy some time soon.  It goes like this.  "You know that promise you keep making that you will prevent the SNP from gaining even a SINGLE SEAT from Labour next May?  I take it you'll be resigning immediately if you don't deliver on that, yeah?"

Because as of this moment Murphy's powers of prophecy aren't looking too hot.  Survation have just released the first full-scale Scottish poll to have been conducted since he became Scottish Labour "leader" (the fieldwork took place between Monday and Thursday) and the message is absolutely identical to the one that Scottish subsamples from GB-wide polls have been sending us over the last few days - namely that the "Murphy bounce" eagerly anticipated by Jim's right-wing admirers north and south of the border has quite simply failed to materialise.  Given that the SNP's Westminster lead in the last Survation poll seemed implausibly high, I had expected them to slip back a bit tonight due to normal sampling variation, but instead there has been a further net swing of about 1% from Labour to the SNP.

Scottish voting intentions for the May 2015 UK general election :

SNP 48.2% (+2.4)
Labour 24.4% (+0.5)
Conservatives 15.9% (-0.8)
Liberal Democrats 5.3% (-0.8)
UKIP 4.0% (-0.8)
Greens 1.1% (-0.4)
BNP 0.4% (-0.3)
SSP 0.1% (-0.1)

On those numbers, Murphy's problems would go a touch further than merely a failure to keep his promise of preventing the SNP from gaining a single seat.  He'd actually be losing no fewer than THIRTY-SEVEN of the forty-one Scottish seats that Labour won at the last general election - and, needless to say, every single one of them would be a gain for the SNP.  Most amusingly of all, Murphy would even lose his own East Renfrewshire seat!

In the real world, the result is unlikely to be quite that good.  It seems reasonable to assume there'll be some kind of swing back to Labour as polling day approaches, if only because they'll have a huge advantage due to lop-sided coverage beamed in by the London media.  It's also possible that Survation are slightly overstating the SNP's lead, even as it stands now.  Although there isn't quite the clear-cut divide that we saw during much of the referendum campaign between Yes-friendly and No-friendly pollsters, Survation do seem to have emerged (at least in their online polls) as one of the most favourable firms for the SNP.  Oddly, though, the most favourable of the lot has been Ipsos-Mori, a pollster that spent the referendum campaign vying with YouGov for the title of most No-friendly firm.  Conversely, the recent polls from Panelbase have shown a somewhat lower (albeit still huge) SNP lead, even though they were for such a long time the only firm that had Yes even within touching distance.

We certainly shouldn't buy into the idea that Survation "must" be overstating the SNP, though.  It's worth pointing out that their methodology is actually much less favourable for the SNP now than it was in their pre-referendum polls, for the simple reason that they have started weighting by recalled referendum vote.  In this poll, 438 people who recall voting Yes were downweighted to count as only 411, while 482 No voters were upweighted to count as 509.  (Those numbers change slightly after undecideds and probable non-voters are stripped out, but the basic pattern remains the same.)  The mind boggles as to how big Labour's deficit would be if the pre-referendum methodology was still in operation.

Whatever we might think about specific weighting procedures and their effect on the size of the SNP lead, one thing about this poll that we should be able to have a reasonable amount of trust in is the trend, because as far as I can see there has been no further methodological change since the last Survation poll.  And it's the trend where the true horror lies for Labour.  The idea that Jim Murphy was going to be some kind of "Messiah" figure now looks unutterably daft.  The Record (which commissioned the poll) are valiantly trying to put a positive spin on the situation by distracting our attention with the results of supplementary questions that asked specifically about Murphy, but in truth even those look pretty grim for Labour.  Just 14% of voters say that they are more likely to vote Labour now that Murphy is "leader", while 18% are less likely to do so.  A quarter of voters say they think Labour will be more successful due to Murphy, compared to 16% who don't - but that's actually a much less important finding, because it will have been influenced more by the media's propaganda efforts to paint Murphy as a political colossus and obvious vote-winner, rather than by respondents' own feelings about the man.

The Record's last throw of the dice is to point to the finding that 28% of Labour voters, 14% of Tory voters and 22% of Liberal Democrat voters agree with the statement : "I am more likely to vote Labour now Jim Murphy is Scottish Labour leader". Suspiciously, though, we are not told how many Labour, Tory and Lib Dem voters agree with the statement "I am less likely to vote Labour now Jim Murphy is Scottish Labour leader", and that part of the datasets hasn't been published yet. Perhaps the most important detail we need to see is how many current SNP voters say they are less likely to vote Labour now that Murphy is in harness - because of course a large chunk of Labour's support from 2010 is now in the SNP column.  Admittedly, we're told that 21% of current SNP voters would "seriously consider voting Labour" - but we aren't told how many current Labour voters would seriously consider voting SNP.  How mysterious.

If you wanted to summarise in a few words why the SNP are doing so well at the moment, it boils down to the fact that they have the support of the vast majority (85%) of people who voted Yes in September, while unionist parties have been unable to prevent a very substantial minority of No voters (29%) from "defecting" to Nicola Sturgeon's party - although in reality a lot of those people will have voted for the SNP in past elections anyway.  It's reasonable to deduce that the SNP need have little fear of suffering localised reverses next year in their No-voting north-east heartlands (and of course that was borne out by a recent council by-election in Aberdeenshire).

Jim Murphy is a dyed-in-the-wool Blairite, which means among other things that he believes in an authoritarian ideology (militarism, detention without trial, and all the rest of it).  It's curious, then, that he's recently spent so much time championing "targetted libertarianism" in a way that he hopes will appeal to a certain group of lapsed Labour voters that he wants to win back - ie. working-class men who he thinks above all else crave the personal freedom to drunkenly shout sectarian abuse at football matches.  I'm just wondering how well a macho offering like that will play with women, especially given that the SNP's new leader is female.  For now, the gender gap still works in Labour's favour (the SNP "only" lead by 46% to 28% among women), but for how much longer?

* * *

Yet more nonsense from Mike Smithson at PB -

"Previous Scottish polls with figures like these have failed to budge the Scottish single seat markets. Last week the SNP was only down as favourite to win 4 seats currently held by LAB. This suggests a lack of confidence on the ground."

No it bloomin' well doesn't. The vast majority of punters at UK betting exchanges are based south of the border and don't have the first clue about what's happening on the ground in Scotland (exactly the same problem that distorted the markets during the referendum). If you wanted to predict an English election using the "wisdom of crowds" principle, you wouldn't ask crowds in Belgium, would you?

"53% of those sampled in Survation Scotland poll "remembered" voting YES in IndyRef"

Heaven only knows where he's getting that from, but it's certainly not from the Survation datasets. In fact, 44% of the unweighted sample recall voting Yes, and 49% recall voting No (the remainder presumably didn't vote or can't remember).

* * *

Survation also have figures for Holyrood...

Constituency voting intentions for the 2016 Scottish Parliament election :

SNP 50.8% (+0.8)
Labour 24.6% (+1.6)
Conservatives 15.1% (+1.0)
Liberal Democrats 4.8% (-1.9)
UKIP 2.2% (-0.9)
Greens 1.5% (-0.8)
BNP 0.3% (-0.5)

Regional list voting intentions for the 2016 Scottish Parliament election :

SNP 39.8% (-0.8)
Labour 23.8% (+3.5)
Conservatives 13.7% (+0.7)
Greens 9.0% (-0.9)
UKIP 7.1% (-0.6)
Liberal Democrats 5.5% (-0.9)
SSP 0.5% (-0.4)
Christians 0.3% (+0.3)
BNP 0.2% (-0.1)

In spite of the slight uptick in Labour support (from a horrendously low base), this poll confirms what other firms have consistently shown - that the biggest threat to the SNP at Holyrood does not come from unionist parties, but instead from the inclination of some of their own voters to drift off to the Greens on the list.  If things remain as they are, that phenomenon would threaten the SNP's chances of a second overall majority.  If the state of play changes, it might even threaten the SNP's chances of remaining the largest single party.  Hopefully a very aggressive two-vote strategy will be put in place to counter this potentially huge problem.

*  *  *

SCOT GOES POP POLL OF POLLS

Barring the improbable appearance of another full-scale Scottish poll between now and Hogmanay, this will be the last Poll of Polls update until 2015.  It's based on the Survation poll, plus five Scottish subsamples from GB-wide polls - four from YouGov and one from Populus.  As always, the new Opinium poll (in which the SNP are flying high) is excluded, because the Scottish subsample result hasn't been published.

Scottish voting intentions for the May 2015 UK general election :

SNP 45.8% (+3.4)
Labour 25.3% (-1.0)
Conservatives 16.5% (+1.0)
Liberal Democrats 5.3% (-1.3)
UKIP 3.7% (+0.1)
Greens 1.9% (-1.7)

(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.)