Friday, February 12, 2021

There's no particular reason to think opinion polls are overestimating the Yes vote

First of all, I have an analysis piece in The National today about the new Savanta ComRes poll which puts the Yes vote on 53%.  You can read it HERE.  

Elsewhere in the same paper, Shona Craven argues that the Yes lead in the polls might be misleading, and asks whether Yes voters are significantly more likely than No voters to join volunteer online polling panels.  To which the answer is: almost certainly yes, but it probably doesn't matter that much, because that's what political weighting is there to correct for.  Most polling firms start by asking respondents how they voted in the 2014 referendum, or in some cases it's already known how they voted due to responses in previous surveys.  If there are too many Yes voters in the sample, they are downweighted accordingly.  For example, in the recent Panelbase poll for the Sunday Times, the 494 Yes voters from 2014 were downweighted to count as only 471, and the 560 No voters were upweighted to count as 583.  If that adjustment hadn't been made, the Yes lead would have been higher than the 52-48 that was reported.  And in many other surveys, the correction has been a lot bigger than that.  

If anything, the greater concern ought to be that an over-correction is going on, and that Yes are actually being underestimated as a result.  That's exactly what happened to the SNP in the 2019 election - the polls underestimated them, because people who voted for them two years earlier were being downweighted by pollsters to a target figure that was too low, due to the disproportionate number of potential SNP voters who stayed at home in 2017.  

Don't get me wrong: I think there's no room at all for complacency about the current polling situation.  But that's not because I think there's any particular reason to suppose the polls are wildly overstating Yes at the moment.  The problem is more that the Yes lead is relatively modest, and public opinion can change in the blink of an eye.

Wednesday, February 10, 2021

Make that TWENTY-ONE in a row: Savanta ComRes poll gives Yes a six-point lead

We now have the first full-scale Scottish poll since the SNP leadership made the idiotic unforced error of sacking their most popular frontbencher just weeks before a crucial election, and we can at least breathe a sigh of relief that Yes still has a lead, and indeed a bit of a cushion.

Should Scotland be an independent country? (Savanta ComRes, 4th-9th February 2021)

Yes 53% (-4)
No 47% (+4)

Depending on your point of view, this is either the twentieth or the twenty-first consecutive poll to show a pro-independence majority - a long unbroken sequence stretching all the way back to the Scot Goes Pop / Panelbase poll of last June.  (It's the twenty-first in a row if you include a Progress Scotland / Survation poll that used a non-standard format.  Personally I think we should be saying it's twenty, but there's such a consensus among Yessers for including the Progress Scotland poll that I've almost given up on trying to make that case.)

When I said that I was worried that the factional behaviour of the leadership in recent days might cost both Yes and the SNP dearly in the polls, what I was really concerned about was that No might return to the lead.  That hasn't happened, but of course the snag here is that Savanta ComRes has recently been one of the most Yes-friendly pollsters.  If, for the sake of argument, YouGov, Survation or even Panelbase were to replicate the four-point drop in Yes support that ComRes are reporting, that would be enough to tip the balance and put No into the lead with those firms.  So the big question is whether ComRes are picking up something real, or whether they're just 'returning to the pack' after a couple of unusually good polls for Yes.  It shouldn't be forgotten that before those two polls, ComRes were reporting very similar figures to other firms - 53% in October (exactly the same as today) and 54% in August.  It doesn't automatically follow, therefore, that other firms will show any drop in Yes support at all - we'll just have to wait and see.

If there has been a genuine swing against Yes, I can think of a number of possible reasons -

1) The sacking of Joanna Cherry.  If it's mainly that, the effect should fade over time, so long as the leadership don't continue to pander to the zealots (for example with further sackings, suspensions or expulsions).  

2) The ongoing fallout from the rift between Nicola Sturgeon and Alex Salmond.

3) The media's determined efforts to portray the vaccine rollout as a success story for the UK Government.  I'd suggest that's the least likely possibility, because the public perception that the UK Government have handled the crisis badly, and that the Scottish Government have handled it well, seems to be pretty much set in stone.

Scottish Parliament constituency ballot voting intentions:

SNP 54% (+1)
Conservatives 23% (+4)
Labour 16% (-3)
Liberal Democrats 6% (-)

Scottish Parliament regional list ballot voting intentions:

SNP 43% (-1)
Conservatives 21% (+5)
Labour 18% (-)
Greens 10% (-1)
Liberal Democrats 6% (-2)

Seats projection (with changes from the 2016 election): SNP 71 (+8), Conservatives 24 (-7), Labour 19 (-5), Greens 11 (+5), Liberal Democrats 4 (-1)

SNP: 71 seats
All other parties combined: 58 seats


Pro-independence parties: 82 seats
Anti-independence parties: 47 seats


So the story of the Holyrood numbers is not one of SNP collapse (because there simply hasn't been one), but of an unexpected Tory recovery - which may have been at the expense of Labour on the constituency vote, but seemingly not on the list.  Why that would have happened isn't entirely clear, because the Tories have been taking a hammering in Scottish polls of late.  I'm wondering if it might turn out to be a freakish finding caused by sampling variation.

I'm slightly bemused by the Scotsman summarising Labour's results as "comfortably around the 18 per cent on both lists".  They're on exactly 18% on one ballot, but they're on 16% on the other.  I'm not sure what person in their right mind would describe 16% as being "around 18%", but each to their own I suppose.

The zealots within the SNP will doubtless seize upon a finding from one of the poll's supplementary questions that 32% of SNP voters back Joanna Cherry's sacking and 13% oppose it - but of course that misses the point entirely.  You'd always expect most of a party's own support base to say they support the decisions of that party's leader, but it's the minority who don't support those decisions you have to worry about.  The SNP simply can't afford to lose, say, one-tenth of their voters just before an election.  Nor can they afford a growing perception that the party is divided - and this poll shows that the proportion of respondents saying the SNP is divided has increased from 39% to 45%.  It's difficult to escape the conclusion that the sacking must be partly responsible for that.

The poll also finds support for GRA reform by a margin of 40% to 26%, which is in stark contrast to other polling in the last couple of years which found overwhelming opposition.  It's highly unlikely that public opinion has turned on its head - the more probable explanation is that this is one of those issues where the answers people give are very much dependent on the way the question is worded.

*  *  *

I became slightly concerned yesterday that my Somerset Stalker hadn't written his 18,573rd obsessional blogpost about me yet - but thankfully that was rectified before the night was out.  Close one.

Early polling straws in the wind after Joanna Cherry's sacking

As people who follow me on Twitter will be aware, I've become increasingly jittery about what upcoming polls will show, both for independence and perhaps more to the point for the SNP - just weeks before a crucial election.  There comes a point where the self-inflicted wounds are severe or prolonged enough that you kind of feel like they're bound to have some kind of effect on the polls.  But we'll have to wait and see - polls often produce counterintuitive results, most obviously the wacky resilience of the Conservative Party in the GB-wide polls in spite of Boris Johnson's catastrophic mishandling of the pandemic.  So if we're very, very lucky, it's not totally impossible that the SNP and Yes will sail through this difficult period relatively unscathed - although I'm certainly not going to predict that.  

Needless to say, there have been no full-scale Scottish polls since Joanna Cherry's sacking, and there have only been two GB-wide polls carried out in their entirety since then - one from YouGov and one from Redfield & Wilton Strategies (a name that always makes me think of Ever Decreasing Circles).  Here are the Scottish subsamples...

YouGov (2nd-3rd February): SNP 53%, Conservatives 23%, Labour 16%, Greens 4%, Liberal Democrats 3%

Redfield & Wilton Strategies (8th February): SNP 43%, Labour 23%, Conservatives 22%, Liberal Democrats 7%, Plaid Cymru 3%, Reform UK 1%

Of course individual subsamples aren't particularly meaningful for all sorts of reasons, although in the absence of any other evidence I always think they're worth keeping at least half an eye on for early warning signs of any dramatic change in the political weather.  YouGov's subsamples can generally be taken a bit more seriously than those from other firms because they seem to be correctly structured and weighted - although the downside in this case is that the YouGov poll (which is excellent for the SNP) was conducted just after the sacking, and if anything the situation inside the SNP has worsened since then.

I've no idea why Plaid Cymru are polling at 3% in Redfield & Wilton's Scottish subsample - it might be an error in the datasets (possibly the Greens' figures placed in the Plaid row), but if it's for real, those votes can perhaps be taken as interchangeable with the SNP's, in which case the numbers are a touch better than they look at first glance.  

So nothing too alarming as of yet - but as I said the other day, if the SNP leadership aren't going to directly undo the mistakes of recent days, they're going to have to find an alternative way of restoring enthusiasm to the membership and the wider Yes movement in the run-up to the election.  We mustn't repeat the errors of 2017 (and to some extent 2016) by demotivating SNP activists and leaving Yessers with nothing concrete to vote for.  We need some really credible specifics about how and when a vote on independence will occur if an SNP majority is secured in the spring.

Monday, February 8, 2021

If the SNP leadership aren't going to reverse the catastrophic mistake of the Great Purge, they urgently need to show signs of having listened to members by offering some red meat on Plan B

Every so often, Stormfront Lite makes a plucky attempt to Anglosplain Scottish politics to us, although just this once - and rather irritatingly - I'm finding it hard to disagree with the overall thrust of David Herdson's piece on Saturday (he's still wrong on a number of details).  Basically he argues that divisions over the trans issue could unexpectedly cost the SNP its chances of an overall majority, and that the fault does not predominantly lie with the rank-and-file but instead with the leadership for giving too much weight to the views of a relatively small minority.  I think that's right - when the zealots called for Joanna Cherry to be sacked, the response should have been to point out that both sides of the debate were represented on the frontbench, and that was exactly as it should be.  The SNP was a broad church, it should have been said, and that was how it intended to stay, rather than letting itself become the hostage of one faction.  Nicola Sturgeon's intervention on social media should have been an even-handed denunciation of both transphobia and any trampling on women's rights, and should have assured both sides of the debate, not just one side, that they were welcomed and valued within the SNP.  

There's nothing wrong with wanting your own viewpoint to be represented at the top - but the problem with the zealots is that they're not satisfied with that, they also want the opposing view to be totally annihilated.  Joanna Cherry's ill-judged sacking did not succeed in appeasing them - it just emboldened them to start preparing the ground for her suspension or expulsion, presumably using the synthetic tool of an absurdly broad definition of the word 'transphobia' which will shortly be written into the party's rulebook.  When I suggested that one solution to the damage that had been done by the sackings was simply to reverse them, it was pointed out to me that although that would reassure a lot of people, it would also cause others to leave the party.  But I do wonder if it might be no bad thing to lay down a marker - to say to people that they're welcome in the party, but that they're not welcome to spend every waking moment trying to drive other good people out.

However, plainly the leadership have no intention of reversing the sackings, so I'd suggest we need something in lieu of that.  Mike Russell argued at the weekend that the danger of division was that it demotivated SNP activists who needed to be wholeheartedly behind the campaign for an SNP majority win in May - but actually the onus has to be on the leadership itself to help solve that problem.  Asking passionate supporters of Joanna Cherry to campaign without reservation for the SNP in these circumstances is a bit like randomly punching a marathon runner in the stomach, and then lecturing him about how he owes it to you and to himself to carry on running for another ten miles.  The problem is particularly acute given that the reshuffle looked like a way of effectively overturning or defying the NEC election results - there can be no clearer way of demonstrating to members that they have no stake in their own party than to treat their votes with contempt.  They indicated that they wanted a change of direction and the leadership's response has been to continue in the original direction at even greater speed.

So we need evidence of the leadership listening, and quickly.  If we can't have Joanna Cherry and people of like mind back on the front bench, probably the best available substitute will be some red meat on Plan B, and it will have to go quite a bit further than Mike Russell's 11-point plan.  We need specifics and dates, and also a Plan C for a plebiscitary election in the event that a legal challenge by the UK Government succeeds.  That, I think, would be the recipe for an enthused membership chapping on doors (or the figurative pandemic equivalent), regardless of any misgivings they have about recent events.

It also goes without saying that there can be no further big concessions to the zealots.  Everyone has their breaking-point, and if MPs or even ordinary members start being suspended or expelled on the basis of bogus allegations of transphobia, people will quite rightly conclude that enough is enough.