Until recently, every poll on Scottish independence since the beginning of 2012 had shown the No campaign in front, usually by a double-digit margin. But that trend ended today with the publication of a new Panelbase survey  putting the Yes camp ahead by 44 to 43 per cent, the first time the nationalist side has led in a poll since August 2011.
The poll was commissioned by the SNP and, as I've noted before, it's always wise to be sceptical of polls published by political parties, principally because the questions asked are often biased in favour of a particular outcome. But on the surface at least, there appear to be no oddities.
Those polled between 23-28 August (the sample size was a respectable 1,043) were asked "There will be a referendum on an independent Scotland on 18th of September 2014. How do you intend to vote in response to the question: Should Scotland be an independent country?" In response, 44 per cent answered "Yes" (up seven points since the last Panelbase poll in July 2013) and 43 per cent answered "No" (down three points), with 13 per cent undecided.
It's a stunning result for the SNP, and entirely at odds with the most recent YouGov poll (carried out a week earlier), which put the No campaign ahead by a record 30 points (59-29). Until other polls are published showing the nationalist side ahead, it's wise to treat survey with caution (lest it prove to be an outlier) but after months of setbacks, the result will be cited by Alex Salmond as proof of his recent claim in the New Statesman that the polls will shift in his favour as the referendum draws closer. He told Jason Cowley: "This is the phoney war. This is not the campaign. I went into an election [for the Scottish Parliament] in 2011 20 points behind in the polls and ended up 15 in front. The real game hasn’t even started. We are just clearing the ground."
In an encouraging precedent for Salmond, Panelbase was the first polling company to put the SNP ahead in the 2011 Scottish parliamentary election. US polling oracle Nate Silver, who recently declared  that there's "virtually no chance that the Yes side will win", is unlikely to be losing any sleep yet, but for the first time in more than two years, Salmond can point to some evidence that the battle is far from over.
Update: Having looked at the full tables  for the survey, it's now clear what might explain the anomalous result. Those polled were first asked whether they thought Scotland could be "a successful, independent country" and whether they trusted the Scottish government or Westminster to take "the best decisions for Scotland". It's likely that both questions nudged people towards supporting independence in the final question. All the more reason, as I said before, to treat the result with caution.