Wednesday, March 23, 2016

Leave campaign receive big boost in landmark ComRes telephone poll

As you'll remember, the Leave campaign took the lead in a telephone poll last week.  That was a genuinely sensational result, because all previous telephone polls had put Remain ahead by either a substantial or enormous margin.  The snag was that the new poll was carried out by ORB, who hadn't previously conducted an EU referendum poll by telephone.  So there were no baseline figures to work from, and it wasn't entirely clear whether there was some sort of ORB 'house effect' favouring Leave, or whether Remain's big lead in telephone polls had genuinely evaporated.  The first evidence today from a non-ORB telephone poll suggests that it may be a bit of both.  ComRes still have Remain ahead, but the lead has slumped to its lowest level to date, and Leave now appear to be firmly within striking distance.

Should the United Kingdom remain a member of the European Union or leave the European Union?

Remain 48% (-3)
Leave 41% (+2)

There's still a missing piece in the jigsaw, though.  ComRes first reported significant slippage in the Remain lead a good few weeks ago, and yet the only Ipsos-Mori telephone poll since then showed business as usual.  Logically, you'd think that the next Ipsos-Mori poll would be bound to finally start picking up the same trend as ComRes, but if it doesn't we'll be left scratching our heads slightly (and not for the first time).

Online polls have of course been showing much more favourable results for Leave all along.  That wouldn't necessarily stop them showing the same direction of travel as phone polls, but it's extremely unclear whether that's happening at the moment.  ICM appear to have detected a small swing to Leave in recent weeks, but the last batch of YouGov polls (which admittedly were quite a while ago now) suggested that it was Remain that had made some progress. 

And we don't have any post-Brussels polls yet, so for all we know the entire state of play may have been rapidly transformed by yesterday's horrific events.  It's hard to know which side that might benefit - if there's an urge to show solidarity with Belgium and France, that ought to work in Remain's favour, but of course anything that promotes irrational fear of immigration might be expected to boost Leave.  We'll just have to wait and see.

I'll calculate the Poll of Polls update later, but I've no doubt at all that the 50/50 online/telephone average will show the Remain lead at a new all-time low.

UPDATE : In fact, it's pretty extraordinary - the telephone and online averages have more or less converged.  That's very likely to be a freakish finding, because there are only two polls in the telephone sample, and one of them is the ultra-Leave-friendly ORB poll.  But the rules are the rules.


Should the United Kingdom remain a member of the European Union or leave the European Union?


Remain 44.2% (-1.6)
Leave 42.6% (+2.9)


Remain 40.9% (-0.3)
Leave 40.1% (+0.1)


Remain 47.5% (-2.8)
Leave 45.0% (+5.7)

(The Poll of Polls takes account of all polls that were conducted at least partly within the last month. The online average is based on ten polls - four from YouGov, four from ICM, one from ORB and one from TNS. The telephone average is based on two polls - one from ComRes and one from ORB.)

Monday, March 21, 2016

Bungling Sunday Express left red-faced after catastrophically misinterpreting research which shows a massive surge in support for independence

As long-term readers of this blog will know, the Scottish Social Attitudes Survey generally shows much lower support for independence than standard opinion polls.  That may be partly because of the data collection method, but the most obvious explanation is the fact that a straight 'Yes/No' question isn't asked - instead, respondents are invited to choose between independence, devolution, and no Scottish Parliament at all.  The reason that's done is to make the results comparable with surveys from previous years, but it inevitably has the effect of steering some respondents towards the "default middle" option of devolution.

Given the difficulty of obtaining a positive result for independence in this format, it's hugely encouraging to see that the latest Scottish Social Attitudes Survey shows support for independence surging by 6% over the last two years, and reaching its highest level ever in the series (and by quite some distance).  That's almost certainly a legacy of the Yes campaign's success.

Independence 39% (+6)
Devolution 49% (-1)
Full London Rule 6% (-1)

As you can see, the number of Don't Knows has also fallen significantly - again, that might be explained by the level of engagement during the referendum, although to be fair the undecideds were unusually numerous in the last survey.

So an all-round good news story for the pro-independence movement.  Spare a thought, however, for poor old Greg Christison of the Sunday Express, who humiliated himself yesterday by going through various contortions of pseudo-logic to try to portray the survey findings as somehow dreadful for both independence and the SNP.  His basic approach was to strip out the Don't Knows and lump together supporters of devolution and full London rule, and then treat them as if they were all directly equivalent to "No voters" in a straight Yes/No opinion poll on independence.  This voodoo method produces a split of Yes 42%, No 58%.  What Greg mysteriously fails to tell you, however, is that the equivalent figures in the last survey were roughly Yes 37%, No 63% - which means there has been an enormous 5% swing in favour of independence over the last two years.  The best result for independence appears to have been eleven years ago, when the equivalent figures were roughly Yes 38%, No 62%.  So, quite literally, the new survey shows that support for independence is higher than ever (or at least higher than at any point since the series started in 1999).

Greg's readers, however, may have taken away a somewhat different story.  They were presented with a fatuous apples-and-oranges comparison between the multi-option SSAS question and straight Yes/No public opinion polls, and were invited to believe that this means that support for independence has fallen to its lowest level since August 2014.  What's particularly nonsensical about this claim, of course, is that the SSAS figures are always way out-of-date by the time we see them - some of the fieldwork dates back to last summer, and all of it was completed by January.  That means we already have several independence polls that were conducted more recently than the SSAS - and the most recent one with a 'real world' methodology (either telephone or face-to-face) produced figures of Yes 52%, No 48%.  That was the Ipsos-Mori poll conducted in early February - just six or seven weeks ago.

I particularly love the bit where Greg hilariously claims that the supposedly awful results for independence in the SSAS came about in spite of the "selection process favouring the SNP", who had 543 of their supporters in the sample.  Er, Greg, I don't know how to break the news to you, but those 543 people made up just 42% of the sample of 1288.  That's considerably less than the 50% of the electorate who actually voted SNP in last year's general election.  As you'd expect, there are some people in the sample who don't identify with any party at all, but the breakdown between the parties looks pretty much bang in line with how people voted.  So no get-out clause for the unionists there.

My suggestion to Greg would be two weeks in a Buddhist retreat.  There's no better way of reminding oneself that the universe is actually a pretty simple sort of place, and that (outside the offices of the Sunday Express) two plus two does not equal eight hundred and seventy-three.

Explanatory note : As a fond tribute to the mainstream media's restrained take on the GERS report, Scot Goes Pop headlines will contain 50% added hysteria for an indefinite period.