Polls are a bit like buses, you wait four months for one and two come along at the same time. After last weekend’s YouGov, there’s a cheeky Panelbase follow up. Running from the 8th to the 13th, the fieldwork pre-dates all the Brexit drama, which some on Twitter reckon might have shifted the polls. I’m afraid we’ll have to wait and see if that really is the case.
This is also the first proper Panelbase poll since the 2016 election, because for some bizarre reason, the Sunday Times didn’t seem to think they needed to ask how people would use their regional vote. It’s fine, guys, it’s only 56 seats in parliament, they don’t matter, right? Fortunately, they’ve come to their senses and commissioned a proper poll this time.
Display format;
- Party – Vote% (Change vs last poll by agency / Change vs last election)
Holyrood Voting Intention and Projection (Tracker)
Constituency;
- SNP – 40% (nc / -7)
- Conservative – 28% (nc / +6)
- Labour – 22% (-1 / -1)
- Lib Dem – 6% (nc / -2)
- Green – 2% (-1 / +1)
Regional;
- SNP – 36% (n/a / -6)
- Conservative – 26% (n/a / +3)
- Labour – 23% (n/a / +4)
- Green – 7% (n/a / nc)
- Lib Dem – 6% (n/a / +1)
- UKIP 1% (n/a / -1)
- SSP – 0% (n/a / -1)
Having two polls in quick succession is a blessing and a curse. They can confirm the current state of play just as well as they can show that polling isn’t an exact science. Resist the temptation to view this poll as “X party has gone up/down since last week” – the difference is far, far more likely to be just bouncing around the margin of error.
The SNP are static compared to the last Panelbase constituency poll, and identical to the YouGov, but may take some heart from the fact that’s the highest they’ve polled on the list vote since March last year. For Labour this is a marginally better poll than YouGov was, whilst the Greens and Lib Dems come off worse. In fact, this is the lowest both parties have polled since October for the list vote. It’s much of a muchness for the Conservatives versus YouGov, keeping the slightly up on 2016 trend going.
If you remember last week YouGov had the SSP on a rather remarkable 3%. They don’t even show up in Panelbase. Panelbase don’t seem to prompt for the SSP, which may account for some of the difference, but even then only one solitary soul proclaimed themselves to be “other”.
If we project that into seats it might look like;
- SNP – 56 (n/a / -7)
- Conservative – 33 (n/a / +2)
- Labour – 28 (n/a / +4)
- Green – 7 (n/a / +1)
- Lib Dem – 5 (n/a / nc)
The map of constituencies is identical to last week’s YouGov, with that big sweep up from Perthshire through Angus, Aberdeenshire and into Moray projected Tory blue. The list is another 66-63 pro-Union majority, but with a few intra-constitutional shuffles. Though the Greens would be up one seat on 2016, this has them on 2 less than YouGov, whilst the SNP are 2 better than they were with YouGov but still down on 2016. The Lib Dems would be static on 2016, 2 less than YouGov, with Labour making up the difference.
As ever, bearing in mind that projections are not an exact science, it’s worth showing the projected winning margin in each constituency to show how close some competitions are. I’m trying another new little thing and pairing this with a map showing the projected runner up in each constituency. Between the above projection map and these two, you should be able to get a sense of how close the race is in each constituency and between which parties.
Westminster Voting Intention (Tracker)
SNP – 38% (+2 / +1)
Conservative – 27% (-1 / -2)
Labour – 27% (nc / nc)
Lib Dem – 6% (nc / -1)
Green – 2% (nc / +2)
Again, not so very different from what YouGov found, though with the SNP a little bit lower and Labour a fair bit higher, putting them neck and neck with the Tories. In the initial data release, they didn’t actually have these available, hence these being edited in two weeks later – full tables are here.
Independence Voting Intention (Tracker)
- Yes – 41% (nc)
- No – 53% (nc)
- Don’t Know – 6% (nc)
Unlike YouGov, Panelbase reported the Don’t Knows for the independence question. Given it’s otherwise a binary question, it’s worth keeping tabs on that figure too. If you exclude them though, everything is still sitting neatly at the status quo since 2014.
- Yes – 44% (nc / -1)
- No – 56% (nc / +1)
That’s it for this week. Tune in again in four months time for another “no polls for ages then two at once” situation.
