Poll Analysis: Panelbase 6th – 10th of September 2021

Early September turned out to be a busy few days for Scottish polling, with an Opinium finishing on the 8th, ComRes on the 9th, and this Panelbase (for the Sunday Times) on the 10th. Not only does that give us quite a good spread of poll findings, but this is Panelbase’s second poll since May’s election, so we’re now starting to be able to compare between polls as well.

Although this wrapped up on the 10th of September, it doesn’t seem to have been widely publicised, which is why this piece is so late. The only reason I spotted it at all was that the Independence side of it had been picked up over on What Scotland Thinks almost immediately, but even they didn’t seem to have initially seen the other parts of it, and thus I didn’t go looking deeper initially. It’s pure chance a check back later for further detail revealed it was a full poll.

The previous Panelbase covered the 16th – 24th of June. Changes are shown as (vs that poll / vs last election).

Regional Vote

Not a lot of movement over the summer, basically. The SNP, Conservatives and Greens are all precisely where they were in the last poll, though for the SNP that means down a bit on May, and for the Greens up marginally. Labour gain a point to put them back where they were at the election, whilst what passes for the biggest change would be for the Lib Dems, up comparatively substantially versus both the last poll and the election.

I noted in another recent analysis piece that the Lib Dems do often hit the 7-8% mark in between elections then fail to live up to it, but this is a second pollster in as many days thinking they are on the up. I wouldn’t be counting any chickens yet if I were them, but I might be beginning to feel a little bit cheerier about future prospects.

Be aware that most of the between-poll swing here is away from “Others”. If you’re looking for Alba, Panelbase don’t appear to be prompting for them, or anyone else – though even if they were, I’d expect very low shares indeed.

Constituency Vote

A sort of mirror image between the SNP and Conservatives here as both drop 2 points, but for the SNP that’s since May’s election and static on the previous Panelbase, whereas the Conservatives are level with May but down on the poll. Labour meanwhile suffering on both measures, down significantly compared to their election share and a little since last poll.

Compared to their list vote figure there’s nothing to excite the Lib Dems here, as gaining a point since June’s poll just brings them to what they won in May. For whatever reason, Panelbase are also finding relatively chunky shares for the Greens on this ballot, who are also up very slightly from the last poll.

Seat Projection

Projecting that into seats might give us something like this:

Please see this page for how projections work and important caveats.

Whereas June’s poll suggested a trio of SNP constituency losses, the Conservative drop back to their May share would only tip Ayr, hence the SNP’s net loss of just one seat. That comparatively weighty Lib Dem share suggests they’d almost double their seat share, giving them a slightly stronger increase than is indicated for the Greens. Overall however, as you’d expect so soon after the election, this wouldn’t be too different to the Parliament we elected then, with a clear majority for the SNP-Green cooperative government.

The usual story for the SNP of a massive lead, and similarly to the Holyrood constituency vote, there’s a bit of dip here for the Conservatives compared to the last poll. There’s a corresponding increase of a single point for both Labour and the Lib Dems, though the whole shuffle is entirely within margin of error.

I do have to note a small beef with Panelbase for slightly inconsistent Green reporting. Either report them for both the Holyrood constituency and Westminster, or pile them in “Others” for both – don’t mix and match between questions. 

On the big constitutional question there’s a very small uptick for No, but coming from the Don’t Knows rather than Yes respondents. That means that the purely binary split is the same 52:48 as the last poll. If that ratio has haunted your dreams since Brexit, then brace yourself…

The table on my tracker shows 7 polls per page, of which 6 out of the 7 most recent have this same 52:48 lead for the Union. It does therefore appear the electorate has settled somewhere in the vicinity of the cursèd numbers for the time being.

Timing of a Referendum

This is the same question with the same answers as they asked in the last poll, and the overall findings are almost identical. That means whilst there’s a plurality against a referendum in this term of parliament, there’s a narrow majority in favour, albeit with many more of those voters wanting it later in the term.

Likely Timing of Independence

Respondents were also asked “Regardless of whether you support or oppose independence, which of these is closest to your view?” They were then given a list of timescales in which they thought Independence was likely to happen within. I didn’t pull this out of the last poll even though they asked it, but it caught my eye this time as a potential curiosity.

Roughly as many voters think Scotland will become Independent within 5 years as think it won’t for decades (or presumably at all – the lack of “Never” option might skew things very slightly). Indeed, more people have gravitated to either extreme since June’s poll. Naturally, people who support Independence are more likely to think it is coming soon, whilst supporters of the Union tend to think it definitely isn’t – nothing too surprising thus far.

However, despite this same poll finding support for the Union in the lead, 48% of respondents think Independence is likely this decade, against 42% who don’t, and 10% who didn’t hazard a guess. Looking slightly further ahead ahead, a clear majority of around 59% think it’ll happen at some point by the middle of the next decade.

Partly, there appears to be an optimism (or pessimism) gap – whereas only about 5% of Yes-leaning respondents picked the longest two timescales, roughly 18% of No-leaners picked the shorter two. In addition, though the constitutional Don’t Knows are quite a small group, they were almost four to one (41% vs 11%) in selecting the the shortest vs longest timescales.

Basically, to sum this whole constitutional section up, according to this poll Scots currently narrowly support the Union, equally narrowly support having a second Independence referendum within 5 years, and are strongly convinced Independence will happen within 15. Taken together, that may go some way to explaining why the nation continues to feel rather stuck on this issue. 


As ever, the last little bit of analysis concerns those hypothetical and more proportional voting systems that I have a bee in my bonnet about here at BBS. The fact Westminster uses pure FPTP is an affront to democracy, and though Holyrood fares far better, AMS is still deeply imperfect. The examples here simply transpose the poll findings onto more proportional voting systems – the reality is that different systems would of course result in different voter behaviour.

For the moment, although the maps are useful for illustrative purposes, I’m opting just to show these hypotheticals as charts. It’s very time consuming making maps, and for these pure hypotheticals, it’s possibly a bit overkill.

Although this is an improved form of AMS, it’s still not perfect. Despite the fact the three pro-Union parties have a narrow vote share lead, this would still lead to a majority for the cooperative government. However, it does substantially reduce the majority, as the net total of seats between the SNP and Greens is the same both on the ordinary and reformed AMS projection – and since there are more seats in the latter, that means a lower share.

If the system were fully proportional, then we’d get a result that has the constitutional balance the right way round in matching seats to votes. In this case the cooperative government would come a couple of seats short of a majority, giving the pro-Union parties a combined lead of 66 seats to 63. That’s a reversal on the narrow pro-Independence party lead for this model in the last poll.

Scandinavian Style Westminster

Note that there’s no comparator to a normal projection here, as I don’t do pure FPTP projections. The differences with FPTP however would be extremely stark, this giving the SNP a reduced and far more accurate (and democratic) share of seats. They’d still be miles out in front however, though just shy a majority of Scotland’s seats. Compared to the last poll, there’d be a very slight shift of a single seat from the Conservatives to Labour.

If you find this or other Ballot Box Scotland output useful and/or interesting, and you can afford to do so, please consider donating to support my work. I love doing this, but it’s a one-man project and takes a lot of time and effort. All donations, no matter how small, are greatly appreciated and extremely helpful.
(About Donations)