Sunday, May 2, 2021

Sensation as new Panelbase poll shows pro-independence majority

Well, I must admit I didn't see this one coming.  I thought the question in tonight's polls would be whether we'd see an even split, a small No lead, or a larger No lead like the one we saw in the recent ComRes poll.  Instead Yes have stormed back into the lead in the Sunday Times' final Panelbase poll before the election.  

Should Scotland be an independent country? (Panelbase / Sunday Times)

Yes 52% (+3)
No 48% (-3)

The percentage changes are from this blog's own Panelbase poll which I published just two days ago.  With a BMG poll in the Herald on Sunday apparently showing a 50-50 split, I think it's becoming reasonably safe to conclude that the ComRes poll may have been an outlier caused by random sampling variation.  That doesn't necessarily mean that Yes are ahead - it could be that ComRes were on the low side and Panelbase are on the high side.  But it does look like talk of a Yes collapse was massively overblown.

And, I'm afraid, that doesn't just mean overblown by unionists - certain Yessers with a vendetta against Alba seemed to have fallen in love with the idea of a drop in support in independence, because they wanted something to blame on Alex Salmond, even though the media blackout of Alba made it wildly improbable that he could have been responsible.  But as he's apparently so all-powerful, I presume his critics will now be logically consistent and give him full credit for this remarkable Yes comeback tonight.  

Scottish Parliament constituency ballot:

SNP 48% (+3)
Conservatives 21% (+1)
Labour 20% (-2)
Liberal Democrats 7% (-1)
Greens 3% (-1)

Scottish Parliament regional list ballot:

SNP 39% (+3)
Conservatives 22% (+1)
Labour 16% (-2)
Greens 8% (-2)
Liberal Democrats 7% (+1)
Alba 4% (-2)
All for Unity 2% (-)

Seats projection (with changes from 2016 election): SNP 65 (+2), Conservatives 28 (-3), Labour 18 (-6), Greens 9 (+3), Liberal Democrats 6 (+1), Alba 3 (+3)

SNP: 65 seats
All others: 64 seats 


Pro-independence parties: 77 seats (59.7%)
Anti-independence parties: 52 seats (40.3%)


The seats projection above is the official one by John Curtice, commissioned by the Sunday Times to accompany the poll.  Ironically if the same projection model that was used for the Scot Goes Pop / Panelbase poll a few days ago had been applied tonight, Alba would have been projected to be on zero seats.  However, Professor Curtice's projection is a much more realistic representation of what would happen in the real world - in practice it's hard to imagine a party not taking any seats at all on 4% of the national list vote.  As I mentioned the other day, no party in Scottish Parliament history has failed to win a seat at that level of support.  And although Alba's 4% vote share is 2% lower than in previous Panelbase polls, it's still their fourth-highest showing in any poll to be published throughout the campaign.

I still can't find the BMG numbers, but according to the Herald's front page they also show an overall SNP majority.  If so, the two polls in combination suggest a mini-wobble for the independence movement may have passed.  The SNP's constituency vote is back above 2016 levels in the Panelbase poll, so they can start thinking about gaining constituency seats once again, rather than losing seats.

More details and analysis to follow...


  1. So now it's 77. Is there any number that is enough for the media?

  2. The recent wobble will likely have been reversed by the revelations on Boris'"pile 'em high" statements and decor choice. Unlike Alba these have had very prominent place in the media over the last 3-5 working days.

    Something else of this magnitude needs to keep on going.

    Also, it.occurred to me that many people may have already voted by post (I have), and how does this reflect in the figures?

    1. The polling companies often take that into account with a 'have you already voted by post?' question. I think somewhere I've got the full questionnaire from the poll I commissioned - it didn't occur to me to check if there was anything like that, but I will do.

  3. Excellent results! Yes, I did this poll as well as the BMG one. One of them had the 'have you already voted?' question. I can't remember which though. I'm not sure but think it was the BMG.