Scottish poll makes me ask: Has Yes been in the lead all along?

In any case, in common with all of their fellow British Polling Council-affiliated firms, Find Out Now does not change its methodology on the main voting intention questions to suit its client. It conducts an independence question for a pro-independence client in the same way as it would for a Unionist or non-aligned client. So the consistency of the Yes lead it is reporting is very real. It moves the pro-independence campaign into the kind of territory that the former Labour leader John Smith famously described as a "settled will". 

READ MORE: England should stop singing God Save the King to calm Scottish booing, says MP

The Yes lead may be modest, but it doesn't appear to be affected much by transient political events. It's just always there, as a constant. That's hugely significant, because - rightly or wrongly - some influential Yes supporters regard a settled will in favour of independence as a prerequisite for moving towards a decisive democratic event that could see Scotland depart from the UK.

All of this, of course, presupposes that Find Out Now's methodology is broadly right. But that's no more or less ridiculous than assuming that any other polling firm is getting it broadly right. A politician who confidently insists that the Yes vote is below 50% is not basing that assertion on election results or on an equivalent form of indisputable evidence, but instead on their own personal faith that "the pack" of pollsters must be right and "the outliers" must be wrong. The inconvenient reality is, though, that there is a long history of the methodology of different polling companies converging, almost on a "safety in numbers" instinct. That leaves firms who stick to a diverging methodology looking a little lonely, but it doesn't necessarily leave them looking wrong.  Accuracy is not decided by majority vote.  For example, in the 2017 General Election, Survation's poll results were way out of line with most of their competitors - but on election night, it turned out that Survation had been right all along and the others had been wrong.

READ MORE: Yes march in Glasgow to go ahead despite Chain of Freedom clash

The truth is that no-one knows for sure whether there has been a small anti-independence majority over the last couple of years, or a small pro-independence majority. But for as long as there is at least one polling firm saying that the latter is the case (and in fact there are two), it would be foolish not to at least consider the possibility that they might be right. It's actually even harder to dismiss Find Out Now's results than it was a few months ago, because they've recently tweaked their methodology by introducing weighting by recalled 2014 indyref vote, which moves them closer into line with other firms, and yet the Yes lead is stubbornly still there.  

The new poll was commissioned by Independent Voices, who like me is sceptical of the wisdom of weighting poll numbers in accordance with how respondents say they voted in a referendum almost a decade ago. There's a huge danger of false recall distorting the published figures, so quite reasonably Independent Voices has published the poll in two forms - one before indyref weighting is applied, and one after.  It turns out that the weighting reduces the Yes lead by one percentage point, if the comparison is made without excluding "don't knows". That's a lesser effect than might have been feared, but it's still enough to warrant a concern that a weighting used by almost all polling firms may have been subtly deceiving us for years about the true state of play in a race that has been very close to a dead heat.