Dramatic poll boost for independence campaign as SNP open up comfortable lead - and almost HALF want a new indyref within just THREE YEARS

Do you remember, around a year ago, before anyone even suspected that Theresa May was going to call a snap election, there was an Ipsos-Mori telephone poll that showed a 50/50 split on the independence question?  We marvelled at how phone polling, which during the long indyref campaign had been much more No-friendly than online polling, was suddenly so favourable for Yes.  If you've bought into what the mainstream media have been telling you since then, you probably think a lot of water has passed under the bridge, and that everything has changed utterly in the last year.  In fact, as today's new Ipsos-Mori telephone poll demonstrates, hardly anything has changed at all.  Yes and No remain in a statistical tie - meaning that due to the standard margin of error, it's not possible to tell which side is actually in the lead.

Should Scotland be an independent country? (Ipsos-Mori, telephone fieldwork, Don't Knows excluded)

Yes 48% (+1)
No 52% (-1)

The reason why the Yes vote has gone up rather than down is that there was also an Ipsos-Mori poll last May that showed a slight dip in Yes support, albeit within the margin of error.

What I find remarkable is that if Ipsos-Mori are typical, it appears that phone polling remains slightly more Yes-friendly than online polling is.  No online poll since June has had Yes higher than 47%.  I must say that's against my expectations, because there were signs in the run-up to the general election that phone and face-to-face polls were swinging sharply back towards No.  There was one particularly awful Survation phone poll in June that had Yes down to 39%, much worse than anything online polls were suggesting at the time.  Perhaps, as was also the case in the EU referendum, phone polls are more likely to be volatile and to show exaggerated swings in either direction.  Whatever the explanation, we can now for the first time say with confidence that, irrespective of data collection method, the Yes vote recovered from any brief election blip, and is back to roughly where it was in the early spring of 2017.

Unfortunately, as is so often the case, I'm going to have to take the media organisation that commissioned the poll to task for misleading people about the results.  STV's online headline, which presumably will be reflected in TV reporting at 6pm, claims that the poll shows that Scots "don't want Indyref 2".  That is quite simply untrue.  The poll actually shows that 42% support a referendum within the next three years (a crucial caveat), 47% oppose a referendum by 2021, 8% "neither support nor oppose" an early referendum, and 3% don't know.  Even if the Don't Knows are stripped out, that means a slender majority of the population do not actively oppose Indyref 2 being held soon.  If Don't Knows and neutrals are removed, the result is as follows -

Do you support or oppose holding another referendum on independence within the next three years?  (Neutrals/Don't Knows excluded)

Support 47%
Oppose 53%

That's extremely close to being a statistical tie, and is strikingly reminiscent of a number of Panelbase polls that have shown a virtual dead heat on similarly-worded questions.

What's even more objectionable about STV's reporting, though, is the unjustified emphasis it's putting on a supplementary question that was asked only to respondents who are in favour of an early indyref.  Those people were asked to make a binary choice between saying that they want a referendum because of Brexit, or that they want a referendum anyway.  STV have used that to split the referendum supporters roughly in half, and to produce a fantasy figure of "only" 22% who supposedly agree with Nicola Sturgeon's stance that a referendum should be held specifically because of Brexit.  Back in the real world, the half of referendum supporters who want an indyref in any circumstances may very well feel that Brexit strengthens the case just the same - but they weren't given any opportunity to say that.  It's also quite possible that some of the respondents who on balance oppose an early referendum nevertheless feel that the arguments in favour are stronger because of Brexit, but they weren't given an opportunity to say that either.  A much more useful question would have been something like "Has Brexit made you more or less supportive of holding an independence referendum within the next three years?" - and it should have been asked to all respondents, not just some.

Scottish voting intentions for next Westminster election:

SNP 39%
Labour 26%
Conservatives 25%
Liberal Democrats 6%
Greens 4%

There have now been eight full-scale Scottish voting intention polls since the general election.  This is the sixth of those to show that the SNP's vote has slightly increased from the 37% recorded on election day.  All eight have shown that the SNP's lead over the Tories has increased, and five have shown that the SNP's lead over Labour has either increased or remained static.

No Holyrood figures from the poll have been published yet, although I would guess STV may release them tomorrow or very soon.  If so, the pattern of recent polls from other firms would suggest that Labour may be performing slightly less well at Holyrood than at Westminster, which could mean they'll still be languishing in third place.

A technical point that will only be of interest to geeks: you may remember that during the indyref campaign we assumed (but didn't know for sure) that Ipsos-Mori were only contacting telephone respondents by landline, which could have meant they were interviewing a disproportionately small 'c' conservative sample.  They now seem to be conceding that point by noting: "Our sample now includes a small proportion of mobile numbers as well as landline."  I've no idea when exactly they made that adjustment, or how much difference it's making to headline results.