Political earthquake beckons as SNP notch up 29% lead for WESTMINSTER in Ipsos-Mori telephone poll

We now have the third full-scale Scottish poll to have been conducted since the referendum - and I'm not sure any superlative is going to prove quite sufficient this time.

Scottish voting intentions for the May 2015 UK general election :

SNP 52%
Labour 23%
Conservatives 10%
Greens 6%
Liberal Democrats 6%
UKIP 2%

So, Mike Smithson, what was that you were saying yesterday about Alex Salmond being a broken man who would struggle to beat the Lib Dems (ahem) in Gordon? I'm not sure your theory about the anti-independence parties "owning" people who voted No last month is looking too hot either - on the above numbers 13% of voters plumped for No in the referendum but are now planning to vote for a pro-independence party at the general election.

What this poll reminds me of most vividly is a couple of polls that were published towards the end of the Holyrood campaign in 2011. Even then, those polls proved a touch too optimistic for the SNP, and in the entirely different context of a Westminster campaign which still has six months and a bit left to run, it's much less likely that the party will be able to pull off a result of quite this magnitude. (The biggest hurdle is that the broadcasters are seemingly still hellbent on moulding the Scottish election result to their own tastes by means of three rigged leaders' debates from which the SNP are totally excluded.) Nevertheless, this is Ipsos-Mori's best estimate of the current state of play, so we can enjoy it for what it is. It seems to me that the "Berlin Wall" that existed between Westminster and Holyrood elections in the public's perceptions has now vanished. Perhaps that's only a temporary phenomenon brought about by the aftermath of the referendum, but for the time being people regard the two elections as being about exactly the same thing - which party is best able to stand up for Scotland.

All the same, the result of this poll is a bit counter-intuitive. Scottish subsamples from GB-wide polls have been consistently putting the SNP comfortably ahead of Labour since the referendum, but we've been exercising a degree of caution with those numbers, because there was always a chance that the SNP's support was being slightly overestimated due to the inherent shortcomings of subsamples. As it turns out, Ipsos-Mori are suggesting that the subsamples have been underestimating the SNP lead - and by quite some distance. The firm is also suggesting that the Greens are doing a bit better than we previously thought, and that the Tories and UKIP are doing quite a bit worse. It's the latter point that might cause us to place a question mark on this poll, because there's a long history of the right-wing vote in Scotland being underestimated by pollsters in Westminster general elections. Given that this is a telephone poll, could there be a "shy Tory" factor at play, with some affluent and rural voters pretending that they are going to vote SNP out of embarrassment? Even if that is what's going on, it's safe to say that the SNP must still have an extremely handsome lead, but perhaps not quite as high as 29%.

It might be thought that Ipsos-Mori's status as one of the most No-friendly pollsters during the referendum (they had Yes on just 36% at the start of March, which seems absurd in retrospect) would lend greater credibility to an insanely good SNP showing, but strangely they also had a reputation for being one of the most SNP-friendly pollsters in the run-up to the 2010 election. That's a contradiction I can't readily explain, although the fact that it's happening with Ipsos-Mori and no other firm may be due to the fact that they don't weight by past vote recall of any type.

This poll is of course totally irreconcilable with the results of the two other post-referendum full-scale Scottish polls - Survation actually put the SNP a few points behind Labour, while Panelbase had them ahead by just two points. In the case of Survation, the difference can perhaps be put down to the fact that fieldwork was conducted on the day immediately after the referendum, when people hadn't yet had a chance to take stock. With Panelbase, the explanation is more likely to be the controversial decision to use a long-discredited weighting procedure, which led to SNP voters being downweighted sharply. So if I was going to hazard a guess (and it's only a guess), I'd say that the true position is most likely to be somewhere in between the two extremes of Panelbase and Ipsos-Mori - indeed it may well be that averaging the subsamples from GB-wide polls has given us a fairly accurate picture all along.

Nevertheless, polling experts such as John Curtice have been studiously ignoring the subsamples, and have almost been playing an intellectual game in which they invite us all to pretend that the only information we have on voting intentions comes from the out-of-date Survation and Panelbase polls (or even from Opinium polls conducted before the referendum!). So the biggest impact of Ipsos-Mori's intervention is that the "official narrative" of where we stand is going to change instantly and radically.

The fieldwork for this poll was conducted between the 22nd and 29th of October, which means that it spans three distinct periods - Wednesday to Friday of last week when Johann Lamont was still in harness, a brief period over the weekend when the media were trying to convince anyone who would listen that either Gordon Brown or Jim Murphy might be taking over, and then finally a period of a few days when Murphy looked like the clear frontrunner. So there's certainly no evidence here that the prospect of a Murphy "leadership" is luring anyone back to Labour. Having said that, the YouGov subsamples over the last couple of days have shown a decreasing SNP lead - but without substantial supporting evidence it would be wrong to assume that's anything more than the normal fluctuations you would expect due to the huge margin of error attached to subsamples.

* * *

SCOT GOES POP POLL OF POLLS

Notwithstanding my suspicion that subsamples are providing us with a more accurate picture at the moment, I do of course give far more weight to full-scale polls in the Poll of Polls. The Ipsos-Mori poll makes up well over half of the sample in this latest update, with seven subsamples also taken into account - four from YouGov, one from Ashcroft, one from ComRes and one from Populus.

Scottish voting intentions for the May 2015 UK general election :

SNP 47.2% (+5.5)
Labour 24.4% (-0.3)
Conservatives 12.9% (-4.6)
Liberal Democrats 6.1% (+0.4)
Greens 5.1% (+0.1)
UKIP 3.1% (-0.9)

(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.)