Thursday, April 22, 2021

A total contradiction: the key trends from the new Savanta ComRes poll cannot be reconciled with the new YouGov poll

We're deep into the 'fog of war' stage of the campaign.  "SNP's hopes of a majority continue to fade" say the Scotsman about their new ComRes poll, with 'continue' and 'fade' looking like weasel words given that a YouGov poll on the exact same day shows the SNP on course for a majority.

Scottish Parliament constituency ballot:

SNP 46% (-3) 
Conservative 25% (+2) 
Labour 20% (+2) 
Liberal Democrats 6% (-)

Scottish Parliament regional list ballot:

SNP 38% (-2) 
Conservatives 23% (+2) 
Labour 17% (-1) 
Greens 7% (-2) 
Liberal Democrats 5 (-1) 
Alba 1% (-2)

The Scotsman's write-up of the poll (written, naturally, by Conor Matchett, notorious for the #Matchettgate fake poll scandal) vaguely suggests Alba are on a "similar" level of support to All for Unity, UKIP, the Abolish the Scottish Parliament Party and the Scottish Libertarian Party.  I would guess that probably means those other four parties are also on 1% of the list vote, but that isn't actually spelt out.

The ComRes and YouGov polls are difficult to reconcile in several ways.  YouGov suggested that the Tories are going backwards on the constituency ballot, allowing Labour to draw level with them and the SNP to further extend their huge lead.  ComRes shows the complete opposite, with the Tories increasing their constituency vote to an unusually high 25% and the SNP lead dropping to an unusually low 21 points.  That's the main reason YouGov show the SNP with an overall majority and ComRes do not - which makes it all the more odd that the ComRes spokesman quoted in Matchett's piece claims that the way the SNP could still win a majority is by substantially boosting their own list vote.  Yes, of course that's a theoretically possible way of doing it, but the much more plausible way would be by securing the sort of margin over the Tories on the constituency ballot that the YouGov poll is pointing to.

ComRes and YouGov also contradict each other on the Green list vote, which is a point of great importance for the pro-indy majority.  YouGov have the Greens surging to an unusually high figure, ComRes have the Greens going backwards.  Clearly they can't both be right.  My gut feeling is that YouGov are probably closer to the truth, but we'll have to await more polls to find out.

Seats projection (with changes from 2016 election): SNP 62 (-1), Conservatives 32 (+1), Labour 22 (-2), Greens 8 (+2), Liberal Democrats 5 (-)

The above projection is my own calculation using a predictor website, and differs from the nonsensical projection in Matchett's piece (good luck trying to get his figures to add up to 129).

SNP: 62 seats
All others: 67 seats


Pro-independence parties: 70 seats
Anti-independence parties: 59 seats


*  *  *

You can catch-up with Episode 6 of the Scot Goes Popcast, in which I speak to Alba Party leader Alex Salmond, HERE (with video) or HERE (audio only).  And if you find Scot Goes Pop's coverage of polls helpful and would like it to continue, I'm currently running a fundraiser HERE.


  1. Comres seem to be all over the p!ace. Given Yougov seem to suggest no significant change from previous polls I think this one needs to go into the "treat with caution" box.

  2. You can still get 2.14 on Betfair on the SNP falling short. Anything above 2 is good value, imo.

  3. Looking at the List vote shares published it would appear that all "Others" account for 9%.

    I find that difficult to believe. And, for me, this casts doubt on the vote shares allocated to the specified parties.

  4. In my averages, I have the SNP constituency vote pretty much flat since mid march at just shy of 50%.

    Since then, they've lost some regional share to the Greens, quite likely inspired by Salmond's 'Second vote Green for a supermajority!' call.

    Alba burst onto the scene at the same time to carve off 3% from the SNP mainly, but haven't moved since (2.8% today). So far not looking likely that they'll take seats. Maybe Salmond himself, but without regional polling, there's no way of knowing.

  5. A lot of the changes between the parties are within the normal margin of error. May not be much of a change by election day. Postal votes have been cast and if the SNP are at 45% or over on the constituency I shall remain content.

  6. The polls have been a mess since feb where the SNP will gain a majority but then 2 other polls say they won't?. The Indy polls are even worse since Feb they've went from a constant 56% Yes lead to a messy 52% ~ 38% Yes level?. The fact that there already on course for a Indy majority really makes hard to take these polls seriously.

    I sometimes wonder If there hoping how gullible people online can get since most of Reddit. Seem to take the current Yes level at face value when showed them being shady they refuse to read or reply.

  7. Regional changes in support since 2016 might mean the SNP and Conservatives swap a constituency or two and Labour will be vulnerable in Dunbarton and East Lothian. However from the looks of polling I feel overall we are heading for a very similar result to last time give or take a handful of seats.

  8. 'Weasel Words'. A perfect review of The Scotsman.

  9. James, any chance you can provide some context about Matchett and the fake poll scandal? I'm not aware of the circumstances of that one.

    1. Pretty simple. He and they published a fake poll. They made a big deal about how shocking it was. People questioned it, including James. They insisted they were right . They weren't . Then they did the non apology apology thing.

    2. Thanks Bill, appreciate the response.