Have ICM cracked the turnout question?
Julian Glover has just tweeted that the new ICM/Guardian poll has (unadjusted) No 44%, Yes 33%, DK 23%, giving an (adjusted) lead of NO 58% vs YES 42%.
This is a massive shift in the polling numbers from the pollster that many (including OGH Mike Smithson) regard as a gold-standard pollster.
The Guardian write-up can be found here: key features are that the December ICM/Guardian poll had YES 6-points ahead, before NO drew level in February. ICM is a telephone pollster, which gives us a contrast with the picture from YouGov which uses an online panel for its tracker.
Also interestingly, ICM has polled the UK – normally, voting intention surveys are limited to Great Britain, so it may be that Northern Ireland (with its experience of STV elections) is bucking the pro-AV inclinations of Wales & Scotland (which use Additional Member Systems for their national devolved legislatures).
The corresponding voting intention figures are Lab (37%), Con (35%), LD (15%), PC/SNP (5%), UKIP (3%), Green (3%) – the high figure (13%) for Others (in Nationalist parties) could be explained by the fact that differential turnout will mean greater voting intention in Wales and Scotland than in England, where not all local councils have elections.
Mike is at the PB.com party near Liverpool Street, so will give his thoughts later on this evening. I hope to see some of you there shortly.