Ipsos-Mori subsample : SNP hold huge lead in Westminster voting intentions

I'm a couple of days late with this, but the latest GB-wide Ipsos-Mori poll of Westminster voting intentions contains very good news in its Scottish subsample -

SNP 41% (+7)
Labour 27% (-8)
Conservatives 25% (+7)
UKIP 4% (+4)
Liberal Democrats 3% (-6)

Of course individual subsamples are notoriously unreliable, so why does this matter? I think the main significance is that it confirms yet again that YouGov have become severe outliers in their Scottish sampling, calling into question their methodology. YouGov are the only major pollster that conduct a daily poll, and they frequently show thoroughly implausible figures for Scotland, with the Conservatives even occasionally finding themselves ahead of the SNP. It must have been months ago that the SNP were last in the outright lead in the YouGov daily subsample (admittedly I may have missed something when I was on holiday in late September), whereas other pollsters that only conduct polls once every few weeks have time and again shown the SNP in the lead. That tells its own story.

It really is time for Kellner and co to go back to the drawing board with their whole approach to polling in Scotland - independence referendum polling, Holyrood polling, Westminster polling, the lot. They ought to start by examining their practice of weighting all their Scottish figures by recalled Westminster vote, rather than by recalled Holyrood vote. It seems highly likely that some people who voted SNP for Holyrood in 2011 are falsely recalling that they also voted SNP in the 2010 Westminster election, which would inevitably lead to the SNP's polling figures being wrongly scaled down by YouGov's weighting.

Apart from anything else, the dubious figures YouGov are reporting for Scotland must be having a marginal distorting impact on their overall figures for Great Britain. That, at least, ought to be sufficient to spur a London-centric organisation into long-overdue action.