A new Guardian/ICM poll has given the No to AV campaign a 16-point lead. Among those who say they are likely to vote in the 5 May referendum and have made up their minds, the poll shows 58 per cent saying No and 42 per cent saying Yes.
Among all respondents, 44 per cent back No and 33 per cent Yes, with 23 per cent saying they don't know.
This is a huge leap from the last Guardian/ICM poll in February which put the two sides neck-and-neck on the same measure. In the equivalent poll in December, the Yes campaign was points ahead.
While other recent polls have shown the no campaign starting to pull ahead, none of the other pollsters has shown such a big lead. On Sunday, an Independent on Sunday/ComRes poll gave the No campaign a 6-point lead.
Why could this be? Well, for a start, it is the first poll for two months to use a random telephone sample, rather than an online panel. PoliticalBetting suggests that it may have cracked the turnout question by asking about likelihood to vote in local elections.
This group of voters tends to be older and more likely to vote No – perhaps explaining why the lead for the No camp increases once intention to vote is taken into account. That goes against the accepted wisdom that a higher turnout will mean a better result for the Yes camp.
Interestingly, despite the huge jump (10 points more than the ComRes poll), I haven't seen any commentators dismiss it as an outlier. On this, UK Polling Report says:
This is the biggest lead we've so far seen for the No campaign in a question asking the bare referendum question, but is very much in line with the "No-wards" trend we've seen from other pollsters. The only company still showing YES ahead in recent polling is Angus Reid.
The poll will be relief for David Cameron and his camp: the Times (£) reports today on the chaos that could be unleashed on the coalition in the event of a Yes vote. According to its coverage, options being considered by Tory MPs include forcing an early election so that it has to take place under first-past-the-post. If these polls are to be believed, however, this will be unnecessary.