ComRes, 6th-13th of August 2020

My regular calls for more polling diversity in a landscape dominated by Panelbase have been answered in double this month. Following last week’s YouGov poll, we’ve had a new kid on the block in the form of Savanta ComRes. Although ComRes aren’t a new pollster overall, they are new to Scottish polling. They certainly haven’t done any voting intention polls since 2016 that I’m aware of. In fact, of the 39 Holyrood polls since 2016, this is only the second not to come from our Triumvirate of Panelbase, YouGov and Survation. We last heard from another pollster, BMG, in October 2016.

However, the exciting freshness of this poll does mean there’s nothing to compare with. We won’t know whether the figures here are high or low by the standards of ComRes until they do another few polls, if indeed they do so. It’s also quite a short poll, just being Holyrood and Independence. Note that fieldwork dates overlapped with the YouGov poll, but with a few more days at the end.

(Also note, for anyone checking the Holyrood poll tracker, I haven’t logged the data for this one in there yet as I’m still anticipating there may be further release from Panelbase over the next few days, and am saving myself the bother of making changes twice!)

Display format;

  • Party – Vote% (Change vs last election or referendum)

Holyrood Voting Intention and Projection (Tracker)


  • SNP – 43% (+1)
  • Conservative – 21% (-2)
  • Labour – 16% (-3)
  • Green – 10% (+3)
  • Liberal Democrat – 8% (+3)

Our only comparison being with 2016, this shows gains for the SNP, Greens and Lib Dems. It says a lot about the recent state of SNP polling that 43% looks quite low, despite still being an improvement on 2016 and just over twice the Conservative figure. Much of that difference is likely to be made up by the fact this is the highest Green figure any pollster has found since April 2019.

Lib Dem and Labour figures are also slightly rosier than recent YouGov and Panelbase findings. The Conservatives are about evens, but that means a decrease since the last election.


  • SNP – 51% (+4)
  • Conservative – 24% (+2)
  • Labour – 17% (-6)
  • Liberal Democrat – 6% (-2)

Interestingly, it’s slightly different for the Constituency vote, with the Conservatives up compared to 2016 and both Labour and Lib Dems down. As with the list figure, other recent highs can give a sense the SNP are “only” on 51% here, although that’s an absolute majority of the vote. With this first poll, ComRes appear to have fallen right in the middle of Panelbase (5% between SNP constituency and list vote) and YouGov (10% between votes) when it comes to the SNP’s gap on each vote, with 8% difference between the two ballots.

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

  • SNP – 67 (+4)
  • Conservative – 26 (-5)
  • Labour – 18 (-6)
  • Green – 10 (+4)
  • Liberal Democrat – 8 (+3)

Yet another poll with a clear SNP majority, though not as large as with other agencies. As the Greens and Lib Dems were up on the all-important list vote, they both also substantially increase their haul of seats. Thanks to that Constituency vote bump the Conservatives are (very narrowly, in some cases) projected to hold all 7 of their current constituencies, but losing list votes contributes to them dropping seats overall.

Constitutional (Independence Tracker)

Usual Independence question:

  • Yes – 49%
  • No – 42%
  • Don’t Know – 9%

Unsurprisingly, this is in line with what other pollsters have found, with a Yes lead over No. This poll comes close to hitting the 50% mark before the exclusion of Don’t Knows. If we take those out, we’re left with:

  • Yes – 54% (+9)
  • No – 46% (-9)

Very nearly flipping the 2014 referendum result. We’re now at the point where Yes has averaged a lead in the polls for six whole months (the trend, per my arbitrary “last 5 polls” measure, started on the 14th of February). If this was uniform swing versus 2014, it’d give 21 councils that lean Yes, and 11 that lean No.


Finally, we’re onto those bits and pieces I like to do to illustrate more proportional election scenarios. Starting with the BBS-favoured “Scandinavian Style” electoral system (changes vs AMS projection):

  • SNP – 56 (-11)
  • Conservative – 28 (+2)
  • Labour – 21 (+3)
  • Green – 13 (+3)
  • Liberal Democrat – 11 (+3)

As ever, the main effect is to correct the SNP’s over-representation. With just 43% of the list vote, they shouldn’t get a majority proportionally speaking, and this model ensures that and sees every other party better off than AMS. However, there’d still be a reasonably substantial lead for pro-Independence MSPs over their pro-Union counterparts, at 69 vs 60.

No Westminster figures in this poll, so the last bit is to show a range of alternative systems that further illustrate how lacking AMS is when it comes to accurately reflecting the views of voters.