Panelbase, 28th Sept-4th Oct 2018

For the second post in as many days on the burst of polling we got on Sunday, it’s Panelbase’s turn. Not only did this poll release on the same day as Survation’s, the fieldwork overlaps almost perfectly, with just a couple of extra days at the end. That’ll make it rather useful for comparative purposes.

We had two Panelbase polls in quick succession in June, but data from the second came out in such a piecemeal manner I totally missed the publishing of full tables, so I’m afraid there’s no link back to a projection from the 27th of June. However, all figures for change do relate to that poll.

Display format;

  • Party – Vote% (Change vs last poll by agency / Change vs last election)

Holyrood Voting Intention and Projection (Tracker)


  • SNP – 41% (nc / -6)
  • Conservative – 26% (-1 / +4)
  • Labour – 21% (-2 / -2)
  • Lib Dem – 6% (nc / -2)
  • Green – 3% (+1 / +2)


  • SNP – 35% (-1 / -7)
  • Conservative – 26% (-1 / +3)
  • Labour – 20% (-3 / +1)
  • Green – 7% (+1 / nc)
  • Lib Dem – 8% (+1 / +3)
  • UKIP – 2% (+1 / nc)

Much like with it’s Survation counterpart, this poll isn’t showing a great deal of movement compared to a few months ago. Labour see the biggest shift on both votes, which does cost them a couple of seats.

Comparing directly with Survation, UKIP barely register and the post-2016 norm of the Conservatives having a quite significant lead over Labour re-asserts itself. YouGov, who are the only other vaguely regular pollster, have consistently shown that as well.

Likewise for the Greens, Panelbase has tended to show lower figures than either of the other main pollsters, hence a static showing in terms of seats. On the other hand, it’s tended to sit in between the two for the Lib Dems. The combination of those two tips the Lib Dems back into fourth here.

Bear in mind though that this is only the third Panelbase poll since 2016, whereas we’ve had 7 from Survation and 6 from YouGov.

If we project that into seats it might look like;

  • SNP – 55 (nc / -8)
  • Conservative – 34 (-1 / +3)
  • Labour – 26 (-2 / +2)
  • Lib Dem – 8 (+1 / +3)
  • Green – 6 (+2 / nc)

In contrast to the poll we looked at yesterday, this would come out with a pro-Union majority in parliament, which is again more in keeping with general trends.

I feel like a broken record always bringing up overhang at this point, but it’s relevant. Note the difference a 2% dip in the SNP’s constituency vote with an equal 2% increase in the Conservative’s makes – it flips half a dozen constituencies. That eats away at the SNP’s overall number of seats, as they are overhanging so heavily in the constituencies that they can’t make the difference back up in list seats, even though 35% is on the higher end of their recent polling on that vote.

Just to really drive home that overhang point, if you equalised the SNP’s constituency and list vote at 41% (say, for ease, by stripping 6% off the Greens), they’d still only come out at 59 seats. The party simply isn’t in majority winning territory at present.

As a final note on constituencies, what we’re consistently seeing is that with Labour pretty much stagnant, the only serious contests are between the SNP and the Conservatives, despite Labour being 2nd in many more constituencies. Basically, there are much smaller SNP-Con margins in those rural seats that flip (or don’t) in these projections than the SNP-Lab margins across the Central Belt.

Westminster Voting Intention (Tracker)

  • SNP – 38% (nc / +1)
  • Conservative – 27% (nc / -2)
  • Labour – 24% (-1 / -3)
  • Lib Dem – 6% (-1 / -1)
  • Green – 2% (nc / +2)
  • UKIP – 2% (+2 / +2)

Pretty much hee-haw movement at Westminster as well according to Panelbase, though given some of the margins Labour won by last year, even a 3% dip might put some seats at risk.

Independence Voting Intention (Tracker)

  • Yes – 41% (nc)
  • No – 52% (-1)
  • Don’t Know – 7% (+1)

Once Don’t Knows are excluded;

  • Yes – 44% (nc / -1)
  • No – 56% (nc / +1)

Note: Change is with poll on 13th of June, as 27th of June poll did not ask a standard independence question.

The constitutional question, static, you say? Who’d have guessed? Fortunately, there are a few other interesting independence questions, so it isn’t complete boredom on this front.

Regardless of view on independence, closest to your view on timescale;

  • Likely to become independent within 5-10 years – 29%
  • 10-15 years – 19%
  • 20-30 years – 10%
  • Don’t think it’s likely – 30%
  • Don’t now – 12%

When should another independence referendum be held;

  • During Brexit negotiations -19%
  • After Brexit negotiations – 28%
  • Shouldn’t be one in next few years – 53%

If there’s a No Deal Brexit;

  • Yes – 40%
  • No – 44%
  • Don’t Know – 15%

No Deal Brexit, excluding Don’t Knows;

  • Yes – 48%
  • No – 52%

The prospect of a No Deal isn’t sending people rushing towards independence in the way many supporters hoped it might, but it is making a substantial chunk of those who’d otherwise support the Union unsure how they’d vote. A reason for both sides of the constitutional divide to be a little cautious, here.

Beyond that though, it’s notable that although there’s a majority against holding another referendum in the foreseeable future, a larger majority reckon Scotland will be independent within 30 years.

Going against my normal behaviour here and dipping into subsamples, it’s not surprising that 82% of 2016 SNP voters think it’s inevitable, on one timescale or another. What’s more surprising may be that a majority of Labour (58%) and Lib Dem (51%) also thought so. Conservatives are unsurprisingly much less likely to agree (26%) – but even then, they don’t have the same unwavering confidence of SNP voters, with only 67% certain it’ll never happen.

Brexit Voting Intention

  • Remain – 59%
  • Leave – 35%
  • Don’t Know – 6%

Once Don’t Knows are excluded;

  • Remain – 63% (+1)
  • Leave – 37% (-1)

As polls Independence, so polls Brexit. Another stonking Remain majority in a Scottish poll, showing Scots are on the whole not budging from their recent referendum votes.