Saturday, March 5, 2016

YouGov eccentrically release four EU referendum polls all at once

Perish the thought that anyone could ever call YouGov "secretive".  Until today, we had assumed that their last four EU referendum polls showed Leave in the lead, but now we learn that they've been quietly polling more recently without releasing the figures - and in all of the last four polls, Remain have been ahead, by anything between two and five points.  That's too consistent a change to be easily explained away by sampling variation, so it seems there has been a genuine and potentially significant swing to Remain among the YouGov polling panel as campaigning has got underway in earnest.

Strangely, though, this can't be taken as an indication that online polling from other firms will necessarily show the same trend.  The fieldwork for the first two of the four newly-released polls either preceded or overlapped with the fieldwork for the most recent ICM online poll, which not only failed to show a swing to Remain, but was actually the joint second-best poll for Leave that ICM have produced.  Perhaps that poll will turn out to be a red herring, but it's at least possible that YouGov and ICM are parting company on the trend in the same way that Ipsos-Mori and ComRes have done recently on the telephone side of the divide.

As you'd expect, the release of four new polls in one go has had a telling impact on the Poll of Polls, with Remain nudging back into the lead on the online average...


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


Remain 45.8% (-0.1)
Leave 38.6% (-0.9)


Remain 40.7% (-0.3)
Leave 39.8% (-1.9)


Remain 50.8% (n/c)
Leave 37.3% (n/c)

(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 thirteen polls - six from YouGov, four from ICM, one from ORB, one from BMG and one from TNS. The telephone average is based on four polls - two from ComRes, one from Ipsos-Mori and one from Survation.)

Tuesday, March 1, 2016

Dazzling double-header of Holyrood polls suggests SNP support is soaring into the stratosphere

We knew that the monthly TNS poll of Holyrood voting intentions would turn up sooner or later, but today unexpectedly brings word of two polls, as the Record's regular pollster Survation starts up a torrid adulterous affair with the Mail.  So much for "Vows", eh?  (That's a joke, by the way, before I get an email from Survation reminding me that they haven't signed a contract of exclusivity with the Record!)

Let's start with the TNS poll, because the numbers are pretty extraordinary -

Constituency ballot :

SNP 60% (+3)
Labour 21% (n/c)
Conservatives 13% (-4)
Liberal Democrats 4% (+1)

Regional list ballot :

SNP 55% (+3)
Labour 21% (+2)
Conservatives 13% (-4)
Greens 6% (n/c)
Liberal Democrats 4% (-2)

It's not unheard of for the SNP to break the 60% barrier in TNS polls, but the only previous times they've done it were in three consecutive polls during their post-election honeymoon last spring and summer.  After that, they slipped down to a consistent range of 56-58%, so today's finding is a bit unexpected.  It could be a freak result caused by a weird sample, but it has to be said that there have been several recent polls from a variety of firms showing at least a small increase in SNP support.  It could be real, therefore, but if so, it's quite hard to pinpoint where the extra support is coming from.  Superficially, this poll would leave you with the impression that it's coming direct from the Tories, but for obvious reasons that's highly unlikely.

It's certainly not clear that it's coming from Labour, because the other big story of the TNS poll is that last month's swing from Labour to Tory (which at the time seemed to corroborate evidence from other pollsters) now looks like a mirage.  Labour have been restored to a commanding 8% lead for second-place, both on the constituency ballot and the all-important list ballot.  Ipsos-Mori have shown a similar trend - a big Tory surge in one poll followed by slippage in the next, so it's not impossible there was a genuine (albeit transitory) spike in support for the Tories, but I can't think of any obvious reason for the timing of it.

I'm now back to my original view that the bookies' very generous odds on Labour finishing second are practically free money.  (The operative word being 'practically', and I'm far too risk-averse to go down that road myself!)

TNS are always the firm that causes the greatest excitement amongst the "tactical voting" brigade, because they consistently show much higher SNP figures than anyone else, leaving the impression that we're heading for a complete "yellow-wash" in the constituency seats..  They're a 'real world' pollster, of course (they conduct interviews face-to-face), so it's perfectly conceivable that they're getting it right and everyone else is getting it wrong.  But the most important reason for a degree of scepticism is that the other 'real world' firm (Ipsos-Mori) have slotted in at the other end of the spectrum, and have shown the SNP as 'low' as 50% - as have Panelbase and YouGov.  If the true figure slipped down to the high 40s, we'd be well into territory where the SNP might need list seats to retain their majority.  On the basis of 45% of the constituency vote last time, they won just 53 of the 73 constituency seats, meaning they required a minimum of 12 list seats for a majority (they got 16).

To return to the point that I raised a number of times in my debate with Tommy Sheridan, the evidence is pretty clear that a so-called "tactical vote" on the list for either RISE or Solidarity is highly likely to be wasted and might as well be thrown in the bin.  (The only possible exception is in the Glasgow electoral region, where it's conceivable that Solidarity has a small concentration of support that the polls are unable to pick up.)  On the headline figures, the TNS poll found just one respondent who is minded to vote Solidarity, and just one respondent who is minded to vote SSP.  That works out as 0.2% of the list vote for both parties.  UKIP, you'll be pleased to hear, find themselves in precisely the same predicament.

Let's turn now to the Survation poll -

Constituency ballot :

SNP 54% (+1)
Labour 21% (-1)
Conservatives 16% (n/c)
Liberal Democrats 5% (-1)

Regional list ballot :

SNP 43% (-2)
Labour 19% (+1)
Conservatives 14% (-1)
Greens 9% (n/c)
Liberal Democrats 7% (+1)
UKIP 6% (n/c)

The fly in the ointment here is that the SNP's list vote has once again slipped below the 44% achieved in the 2011 election.  That could be slightly ominous if it turns out that list seats are required for a majority, although in all probability Survation are overestimating how many SNP constituency voters are likely to switch to the Greens on the list.  John Curtice thinks that the question Survation ask is leading respondents to wrongly think that the list vote is some sort of second preference vote.  If you look at the exact wording used, it's a debatable point -

"Your second vote will be a party list vote to elect representatives from your region of Scotland by a form of proportional representation. If the election were tomorrow, which party would you be most likely to vote for with your second, regional list vote?"

Arguably the second use of the word "second" is a bit redundant.  Whatever the explanation, though, the empirical evidence is that Survation are showing a wildly implausible gap between the SNP's constituency and list share of the vote, which in turn makes it pretty likely that something is going wrong and that the Greens (and probably UKIP as well) are being significantly overestimated on the list.

This poll doesn't show quite as big a gap between Labour and Tory as TNS does, but nevertheless the Labour advantage has widened on the list vote, which adds to the weight of evidence that the Tories are highly unlikely to claim the runner-up spot (barring a total Labour meltdown beyond even anything we've seen thus far).

Minor party news : Survation don't ask about Solidarity or RISE, but the "another party" option claims just 1.1% of the list vote.

By far the most amusing and satisfying detail of this poll, which was commissioned by the Mail, is that it shows a plurality of support for the BBC replacing the network Six O'Clock News with a "Scottish Six" produced in Glasgow and tailored for a Scottish audience.  36% of respondents supported the idea, and only 33% were opposed.  The Mail, of course, have been wrongly telling us for days that the people of Scotland don't want the Scottish Six.

*  *  *


Constituency ballot :

SNP 53.4% (+0.8)
Labour 20.4% (-0.2)
Conservatives 16.4% (-0.8)
Liberal Democrats 5.4% (n/c)

Regional list ballot :

SNP 47.4% (+0.2)
Labour 19.6% (+0.6)
Conservatives 15.8% (-1.0)
Greens 6.4% (n/c)
Liberal Democrats 6.2% (-0.2)

(The Poll of Polls is based on a rolling average of the most recent poll from each of the firms that have reported Scottish Parliament voting intention numbers over the previous three months, and that adhere to British Polling Council rules. At present, there are five - Panelbase, Survation, YouGov, TNS and Ipsos-Mori. Whenever a new poll is published, it replaces the last poll from the same company in the sample.)

Monday, February 29, 2016

"Leave" campaign narrows the gap to just 6.4% in latest EU referendum Poll of Polls

Happy Leap Day, everyone (or whatever it's called). Obviously I've done the prudent thing today and stayed indoors, because I can't be bothered having to fend off the hordes of women proposing marriage. But the fact that we've made it to the end of February means that there only just over two months to go until the Holyrood election, and less than four months to go until the EU referendum. This may not be a Greek-style referendum campaign that lasts only a week, but it certainly seems pretty truncated when you bear in mind how much is at stake.

There have been a couple of online polls published since I last updated the referendum Poll of Polls. ORB have Leave in a 52-48 lead, which is at the upper end of their normal range for Leave, but on the other hand BMG have Remain in a 3-point lead, which is at the upper end of their (relatively narrow) normal range for Remain. Taking the two together, then, it looks like the changes are probably meaningless margin of error 'noise' - but what we really need is another telephone poll, ideally from Ipsos-Mori, because the last two ComRes polls have suggested Remain's big telephone lead may have slipped, even as the online state of play remained fairly steady.

The gap between Remain and Leave on the headline 50/50 online/telephone average is down to just 6.4%, which is the lowest figure since the Poll of Polls started. That's been caused purely by two Remain-friendly telephone polls from January dropping out of the sample.


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


Remain 45.9% (n/c)
Leave 39.5% (+0.7)


Remain 41.0% (+1.0)
Leave 41.7% (+1.0)


Remain 50.8% (-1.0)
Leave 37.3% (+0.5)

(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 nine polls - four from ICM, two from YouGov, one from ORB, one from BMG and one from TNS. The telephone average is based on four polls - two from ComRes, one from Ipsos-Mori and one from Survation.)