From Survation on 14 October 2026, with the 7% Yes lead above, see how they have weighted their sample to take account of how respondents remember voting in 2014 and how they would vote in another referendum:

Note how the sample has been changed from a small Yes lead to a larger No lead in both cases. The percentage changes are, however, single digit.

The result, despite this skewing of the sample is still a 7% Yes lead, the biggest Yes lead since January.

Looking now at the Norstat/Sunday Mail poll which found only a 3% Yes lead, on 23 October 2026, we see instantly the markedly great skewing adjustment of the sample.

The 2014 Yes respondents are reduced/weighted from 388 to 318, 82% compared to Survations’s reduction to 92% and the Yes respondents are increased/weighted to 393 from 322, 122% compared to Survation’s increase of 104%!

The result, despite this skewing of the sample is still a 3% Yes lead.

What about that 12% Yes lead back in January, by Stonehaven for the Sunday Times? They did not weight the sample using the 2014 vote, at all.

What does this tell you? With no weighting/skewing the sample to have it look like that in 2014, the lead for another Indy referendum looks like a 10-12% Yes lead.