An independent Scotland beckons, as Yes campaign close the gap to just 5% in spectacular new Panelbase poll

The third Panelbase referendum poll of 2014 has been released in the last few minutes - and not to put too fine a point on it, it's a cracker.

There will be a referendum on an independent Scotland on the 18th of September. How do you intend to vote in response to the question: Should Scotland be an independent country?

Yes 40% (+3)
No 45% (-2)

With Don't Knows excluded, it works out as -

Yes 47% (+3)
No 53% (-3)

This poll is a landmark moment in practically countless ways. It shows the highest Yes vote to be recorded in any poll conducted by any pollster in referendum year so far. It marks the first time since last summer that Yes have broken through the psychological 40% barrier on the headline figures produced by any pollster (although tellingly the most recent poll from Survation showed Yes edging up to 39%). It shows the lowest No lead to be recorded by any pollster so far this year - and that includes the ICM poll that set the campaign alight back in January.

Most significantly, though, it's the final piece in the jigsaw that confirms beyond any reasonable doubt that the No lead has dipped since September. Until now, Panelbase were the one and only pollster that hadn't found any decrease in the No lead since the publication of the White Paper - which was very unexpected and mysterious given their reputation for producing Yes-friendly results. But now they have finally joined a consensus that even includes the extreme No-friendly outliers Ipsos-Mori. A 5% No lead is well below Panelbase's normal range of 8-13%, and if anything that range should have creeped up rather than down as a result of the recent methodological change. So unless it's an out-and-out rogue poll (which should only happen one time in every twenty) this breakthrough for Yes can't be dismissed as being "margin of error noise".

It's also worth pointing out that these numbers almost (but not quite) represent what American journalists traditionally refer to as a "statistical tie" - meaning a situation in which there is a greater than 5% chance that the side that appears to be in the lead is actually behind, due to the standard margin of error.

Ever since it became apparent that the No lead had fallen by an appreciable amount, I've seen some Yes supporters fretting that the pace of the trend across all pollsters isn't quite sufficient to get us to a 50%+ Yes vote on September 18th. I must say I think that misses the point - any really dramatic changes of opinion, if they occur, are likely to happen during the official campaign period that commences in late May. Look at the overnight Cleggasm, for example, or the SNP's dramatic advance in 2011. Neither of those occurred until voters properly switched on to the choice that lay ahead of them. The significance of a dip in the No lead at this stage is that it demonstrates that the mantra we've heard from unionist journalists and politicians is wrong - public opinion is plainly not set in stone, and if some No voters have been won over already, there is every chance that others will follow if they are exposed to the same arguments over the coming months.

It'll be fascinating to see how (and if) Blair McDougall reacts to this poll. In true Orwellian fashion, the anti-independence camp's embarrassment of a campaign chief treated the last Panelbase poll as an "un-poll" - it simply didn't exist in his eyes, presumably because it was commissioned by the hated SNP. (If you look at his comments about recent polling trends, you'll see they don't make any logical sense unless he is treating the last-but-one Panelbase poll as if it was the most recent one.) Well, this new one was commissioned by Newsnet Scotland, so I wonder if that will be considered better or worse in McDougall-world? He really ought to bear in mind that there's a downside as well as an upside to pretending that bad polls don't exist. Suppose for example that the next Panelbase poll for the Sunday Times shows an 8% No lead - in most people's eyes (and in reality) that would be a 3% increase. But to maintain his cherished fiction, Blair would have to say that the No lead had slumped by 4% since the last Panelbase poll. That would be a rather amusing moment.

* * *

SCOT GOES POP POLL OF POLLS

This update of the Poll of Polls also represents a landmark for the pro-independence campaign, with Yes breaking through the 35% barrier on the headline figures for the first time. The 41.9% Yes vote when Don't Knows are excluded is also the highest recorded so far.

MEAN AVERAGE (not excluding Don't Knows) :

Yes 35.1% (+0.4)
No 48.7% (-0.3)

MEAN AVERAGE (excluding Don't Knows) :

Yes 41.9% (+0.4)
No 58.1% (-0.4)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 42.0% (n/c)
No 58.0% (n/c)

(The Poll of Polls is based on a rolling average of the most recent poll from each of the pollsters that have been active in the referendum campaign, and that adhere to British Polling Council rules. At present, there are seven - YouGov, TNS-BMRB, Angus Reid, Survation, Panelbase, Ipsos-Mori and ICM. Whenever a new poll is published, it replaces the last poll from the same company in the sample. Changes in the Poll of Polls are generally glacial in nature due to the fact that only a small portion of the sample is updated each time.)

Of course these figures are significantly influenced (some would say distorted) by the inclusion of extreme No-friendly outliers Ipsos-Mori in the sample. An alternative way of measuring the current state of play is to look at an average of the most recent figures from each of the four online pollsters that have conducted polls so far this year (Panelbase, YouGov, Survation and ICM), and unsurprisingly that shows a tighter race -

MEAN AVERAGE OF ONLINE POLLSTERS (not excluding Don't Knows) :

Yes 37.8% (+0.8)
No 48.8% (-0.5)

MEAN AVERAGE OF ONLINE POLLSTERS (excluding Don't Knows) :

Yes 43.6% (+0.7)
No 56.4% (-0.7)

MEDIAN AVERAGE OF ONLINE POLLSTERS (excluding Don't Knows) :

Yes 43.9% (+0.4)
No 56.1% (-0.4)