We're still hopelessly stranded in polling Antarctica, as the various firms lick their wounds after their humbling on May 7th. The one full-scale Scottish poll we've had was obviously hugely encouraging for the SNP, but the fieldwork predated the death of Charles Kennedy. That might conceivably be of some importance, due to the disgraceful attempts of some right-wing journalists to exploit the tragedy to damage Alex Salmond. However, we now have a second straw in the wind from a GB-wide poll conducted well into June, and so far there is no hint at all of the SNP surge going into reverse. Yesterday's Ipsos-Mori Scottish subsample has figures of : SNP 56%, Labour 22%, Conservatives 17%, Liberal Democrats 5%, UKIP 1%. As with the ICM subsample that had the SNP in the high 40s, the number of respondents is very low. So the information for June is still extraordinarily limited, but there is certainly no obvious cause for concern so far.
Ipsos-Mori's Scottish sample report a different preference for Labour leader from respondents in the rest of Britain, narrowly plumping for Yvette Cooper over Liz Kendall by a 15% to 12% margin. Andy Burnham, the Britain-wide favourite, trails in a poor third. However, that may well be a freak result caused by the small sample size.
The most significant finding of the poll has only been released today - across Britain, support for remaining a member of the EU has reached a 24-year high, with the equivalent of the Yes campaign in the referendum leading by 61% to 27%. That's based on the long-running tracker question that was asked to half the sample. The other half were asked the actual referendum question, producing an even more dramatic result - 66% Yes, 22% No.
So is it game over before we even start? Answer : no, or at least not until we get some clarity about the reason for the disparity between the results produced by telephone and online pollsters, which is every bit as extreme as we saw in the independence referendum (if not more so). Recent YouGov online polls have shown a Yes lead, but a very modest one. It's easy to jump to the conclusion that telephone polls must be more accurate, or that the truth must be somewhere in between the two extremes - in which case Yes would have a handsome lead. That's not necessarily the case - for example, in the 2008 London mayoral election, YouGov's internet polls comprehensively got the better of Ipsos-Mori's telephone polls.
The other thing that needs to be borne in mind is that the electorate tends to behave in a much more volatile way in referendums than in regular elections. Ipsos-Mori may never have reported a 3-1 No lead in the independence referendum, but they did report a 2-1 lead at one point. We all know how dramatically the race narrowed afterwards.