Panelbase, 24th-26th Mar 2020

Much of our lives may be on hold as the globe grapples with the coronavirus pandemic, but in this time of great uncertainty we can at least rely on there being yet another Panelbase poll. As welcome a respite as this is from writing and thinking about said pandemic, Panelbase polls now make up 40% of the total polling since the last Holyrood election, and a whopping 63% of polls in the past year. I’m begging papers – commission more work from Survation and YouGov! Variety is the spice of polling as well as life. Additionally, from the perspective of someone who reports on every (reputable) Scottish poll that comes out, Panelbase take a frustratingly long time to get their tables online, which also makes covering them more of slog for me over a number of days.

Display format;

Holyrood Voting Intention and Projection (Tracker)


  • SNP – 48% (+1 / +6)
  • Conservative – 26% (+1 / +3)
  • Labour – 13% (-1 / -6)
  • Liberal Democrat – 6% (-1 / +1)
  • Green – 6% (-1 / -1)

January’s Panelbase saw a remarkable SNP surge, and in this poll they’ve even managed to squeeze a bit more on top. These are the two highest list vote figures the SNP have seen since the 2016 election, creeping close to the point of an absolute majority of support. All that said, there’s relatively little movement here, with the Conservatives also creeping up by 1%, and the other three Holyrood parties all dipping by 1%. That’d be mildly disappointing for both the Greens and Lib Dems, but for Labour it’s their second worst list vote poll in the history of the Scottish Parliament.

Note again the complete absence of the Brexit Party. They weren’t prompted for, and only 1% of respondents plumped for “Other”, so we’re probably at the point we can safely say their performance last year was very much a flash in the pan in reaction to the delay in Brexit date.


  • SNP – 51% (+1 / +4)
  • Conservative – 26% (nc / +4)
  • Labour – 14% (nc / -9)
  • Liberal Democrat – 6% (-1 / -2)
  • Green – 3% (nc / +2)

There’s even less movement on the constituency figures, where the SNP again have a little 1% uptick, whilst the Lib Dems drop by as much, and everyone else is static. This is narrowly not the SNP’s best polling figure on this vote, but it is just behind the 52% they saw way back in August 2016, and also matches January’s Survation. Although Labour haven’t declined any further versus the last poll, it remains quite remarkable to see them down a whopping 9% versus their 2016 result.

If we do the usual projection into seats, it might look like;

  • SNP – 70 (+2 / +7)
  • Conservative – 34 (+2 / +3)
  • Labour – 17 (nc / -7)
  • Liberal Democrat – 5 (nc / nc)
  • Green – 3 (-4 / -3)

That’s an extremely clear SNP majority, a seat ahead even of their surprise 2011 result. The previous Panelbase also had them with a majority, but otherwise you have to go back to (all of) the polls between the 2016 Holyrood election and 2017 snap Westminster election to see projected SNP majorities. Conservatives also appear to be continuing to bounce back after Brexit put them in the doldrums last year, recording a joint second best seat haul here. Another interesting little titbit here is that this is the first projection since I started this project where the Lib Dems repeat their 2011 wipeout from mainland constituencies. In reality, I don’t think that’d happen, and this is a reminder of the limitations of projecting individual constituencies from national polls, but it’s certainly a notable outcome from this calculation.

Overall both Labour and the Lib Dems are static in terms of seats despite dropping a bit on the list vote, but it’s the Greens who are the big losers here, dropping from 7 to 3 seats thanks to that 1% decline in vote share. Remember that AMS is regionally, not nationally, proportional. With a well-spread vote share, Greens suffer at the lower end of the polling scale compared to equally supported Lib Dems, who have a more concentrated vote, whereas the reverse becomes true when both parties are even slightly higher.

I always like to do it anyway, but that last point can be illustrated well by running these polling figures through my alternative Scandinavian style electoral system, which is nationally proportional (vs AMS projection);

  • SNP – 62 (-8)
  • Conservative – 34 (nc)
  • Labour – 17 (nc)
  • Green – 8 (+5)
  • Liberal Democrat – 8 (+3)

And with the same overall vote share, the Greens and Lib Dems come out with equal seats. That’s entirely at the SNP’s expense, who come short of a majority of seats here because they don’t have a majority of (parties above 3%) votes.

Now, we all know I love a good electoral system chat, and my calculator has long produced more than just the Scandinavian alternative. I finally got round to squeezing all of that data into a chart that should be relatively easy to read, which compares the actual list vote (remember, that’s the proportional bit of the system) with the number of seats projected to be won under a range of alternative electoral systems. Some are minor changes (replacing the D’Hondt formula with Sainte-Laguë, for example), whilst others are more dramatic. I’ll write a full post about this particular chart at some point soon, but I wanted to include it this time since I’d got it working.

It further drives home that point that whilst smaller parties would be a lot worse off without the proportional element of AMS, they are nonetheless pretty badly damaged by it compared to more proportional alternatives.

Westminster Voting Intention (Tracker – still to be updated!)

This is the first time Panelbase have included Westminster figures since the December election, so changes here therefore obviously only versus that election.

  • SNP – 48% (+3)
  • Conservative – 27% (+2)
  • Labour – 16% (-3)
  • Liberal Democrat – 5% (-5)
  • Green – 3% (+2)

A fair bit going on here, although the changes are mostly relatively small. The SNP are climbing upwards at Westminster as well, as are the Conservatives and the Greens. The latter perhaps is more representative of the Greens’ “natural” level of Westminster support if they stood in every constituency than it is of actual growth, bearing in mind they got 1% from standing in just over a third of seats in December. Somewhat grim reading for both Labour and the Lib Dems here though, with the latter dropping fully half their vote.

Constitutional (Independence Tracker)

Usual Independence Question;

  • No – 47% (+1)
  • Yes – 46% (-3)
  • Don’t Know – 7% (+1)

Excluding Don’t Knows;

  • No – 51% (+3 / -4)
  • Yes – 49% (-3 / +4)

In January we had a tie and two narrow leads for Yes in Independence polling, with the Panelbase coming out the highest with an EU Ref-tastic 52 to 48. This time around, it’s back to the narrow No lead that’s been the norm in polling for ages now. Basically, we’re bouncing around the margin of error, with Scotland not really having any settled sense of which direction it wants to go for the moment.

I haven’t done the usual for fun projection here – if you want to see what a 51:49 No:Yes split looks like, there’s one in the post for the November Panelbase poll.


Since we’re mid-pandemic and the poll included a couple of questions related to it, it makes sense to highlight them here.

Firstly, there was a further Independence related question in there, tying in with coronavirus. Would that crisis make people more likely to support or oppose Independence?

  • Much More Likely to Support – 16%
  • More Likely to Support – 8%
  • No Difference – 60%
  • More Likely to Oppose – 4%
  • Much More Likely to Oppose – 13%

Probably unsurprising to find that most people don’t really figure this situation into their constitutional position – but a substantial minority do. That minority breaks 24% more likely to support independence as a result, versus 16% (rounding) more likely to oppose it. That’s a curious little portion of this poll that I’m not sure what to make of, or indeed if anything should be made of it at the moment.

The other questions related to how good or bad a job leaders were doing handling the crisis. For Prime Minister Boris Johnson, it came out as;

  • Very Good – 17%
  • Good – 30%
  • Neither – 20%
  • Bad – 15%
  • Very Bad – 15%
  • Don’t Know – 2%

So overall, 47% of Scots think the Prime Minister is doing a good job in the current situation, versus 30% who think he’s doing a bad one.

Now, for First Minister Nicola Sturgeon, it was;

  • Very Good – 29%
  • Good – 38%
  • Neither – 18%
  • Bad – 6%
  • Very Bad – 7%
  • Don’t Know – 2%

Again, looking at the overall, that’s 67% of Scots feeling the First Minister is doing a good job facing up to coronavirus, versus 13% who think she’s doing a bad job.