On Wednesday, the Daily Record made a song and dance about a supplementary question in their monthly Survation poll, which purported to show that 51% of the electorate "don't want another independence referendum", and that 32% do. On the face of it, this contradicts a string of polls we've had since September (some quite recently) showing that a clear majority want a second referendum to take place within the next ten years. But that hasn't stopped Jackanory Jim seizing on the results, and claiming that "the opinion polls" show that "the vast majority" don't want another referendum. Only on Planet Murphy can a single aberrant poll be described as "the opinion polls", and 51% be described as "the vast majority".
Anyway, since the Wednesday report I've been waiting patiently to see the datasets for the relevant question, but they mysteriously haven't appeared yet. (Correct me if I'm wrong about that, because Survation's website isn't the most user-friendly, but I've searched as carefully as I can.) What I want to see is whether there is something about the way the question was asked that might explain why it produced such a different result. The Record claim the question was simply : "Do you want another independence referendum?", but given the agenda of that newspaper I'm not prepared to take that on trust. If that really was the question, it may be that people interpreted it as meaning "Do you want another referendum right now, as a direct result of this election?", in which case respondents who want a referendum in a few years' time will have answered No rather than Yes, thus giving a misleading impression. But it's also possible that the question was more complicated than we've been led to believe.
It would be nice (and dare I say appropriate) for us to get answers to these points before polling day, given that this has become such a topic of controversy during the campaign.
[UPDATE, 5.50pm : The datasets have now turned up, and guess what? The Record were cynically misleading their readers. The question included the all-important caveat "before 2020", and it was only one of several to be asked on the same topic. One of the other questions found that 59% of voters want a second independence referendum within the next ten years. Full details can be found HERE.]
Looking at it more broadly, this is a useful illustration of the importance of poll results, and how they're almost becoming "part of the constitution". The SNP would never have been included in the UK-wide leaders' debates if it hadn't been for their wonderful showing in the polls over the autumn and winter, and by extension they would probably now be on course to win fewer seats on Thursday. It looks very likely that polls will also now be the decisive factor in determining the 'legitimacy' of an early referendum, or indeed a referendum at any point.
If you're a member of an online polling panel, it might be worth bearing that in mind. Even if you personally don't think a referendum should take place for, say, another ten years, I'd highly recommend being as positive as humanly possible about the prospect of an early referendum when you respond to polls, because otherwise the likes of Murphy will simply twist the results in an attempt to close down Scotland's democratic options for the foreseeable future.