Mainstream media's narrative lies in TATTERS as SNP vote increases across the board in sensational new Survation poll

"Mike Smithson, calling Mike Smithson...Mike Smithson, come in please."  If you happen to be sitting next to a certain 'impartial election expert' who demonstrated such an inspiring willingness to 'help' the voters of East Dunbartonshire last year, you might want to give him a wee nudge.  Before tonight, there had been six full-scale Scottish polls since the general election, of which four could reasonably be described as good for the SNP - ie. they showed the SNP on a higher share of the Westminster vote than was achieved in June.  Mr Smithson literally pretended those four didn't exist (and even made mind-boggling comments such as "in the complete absence of any Scottish polling...").  He was all over the other two polls like a rash.  But I'm sure that this apparent pattern is just totally coincidental, and that he'll set our minds to rest first thing in the morning with a big splash on Stormfront Lite about tonight's remarkable Survation poll, which shows the SNP making progress at both Westminster and Holyrood level.

Scottish voting intentions for next Westminster election:

SNP 39% (+1)
Labour 27% (-2)
Conservatives 24% (n/c)
Liberal Democrats 7% (n/c)

Scottish Parliament constituency ballot:

SNP 42% (+3)
Labour 25% (-3)
Conservatives 25% (+1)
Liberal Democrats 6% (-1)

Scottish Parliament regional list ballot:

SNP 33% (+1)
Conservatives 23% (+2)
Labour 23% (-1)
Greens 9% (-1)
Liberal Democrats 8% (-2)
UKIP 3% (n/c)

Don't be thrown off the scent by the fact that the Tory vote has proved reasonably stable between the last Survation poll and this one.  The single most important fact about Scottish polling at the moment is that the gap between the SNP and the Tories at Westminster level has undoubtedly widened since the general election, meaning there is every chance that some of the Tory gains last June will be reversed if there is another general election any time soon.  If this particular poll is correct, the SNP lead over the Tories has essentially doubled from eight points in June to fifteen points now, but no post-election poll has had the gap at less than twelve points.

By contrast, although this poll suggests that the SNP lead over Labour has also increased since the election, it's harder to be sure whether that's what has really happened.  Three of the seven post-election Scottish polls have shown Labour making minor inroads into the SNP's lead, so there may have been a small swing in either direction.  Because the SNP and Labour both hold seats with wafer-thin margins over each other, either party would be equally justified in feeling optimistic about making significant gains in SNP-Labour battleground areas.  But obviously if tonight's poll is exactly right, the SNP would be in luck and would take seats from both the Tories and Labour, increasing its tally of Westminster seats from 35 to well over 40.

Now, I know some people will caution that the SNP's improvement in this poll could easily be margin of error 'noise', and that's undoubtedly true.  The 3% swing from Labour to SNP on the Holyrood constituency vote looks more dramatic than changes elsewhere in the poll, but even that could potentially be explained by margin of error effects.  But in a sense that misses the crucial point - what really matters about this poll is that the SNP's vote hasn't fallen.  You might recall that there was a YouGov poll that took us by surprise recently by showing a sharp 4-point drop in the SNP's vote to 36% - the first time the party had fallen below the 37% share recorded at the general election.  (The aforementioned Mr Smithson wasted no time in painting that as an unmitigated disaster for Nicola Sturgeon.)  If Survation had shown something similar tonight, it would have looked very likely that YouGov had been the first to detect a genuine new trend.  As it is, it now looks more probable that the changes reported by YouGov were just a margin of error illusion, and that the SNP's Westminster vote has held steady - perhaps somewhere in the high 30s.

Oddly enough, Labour have more or less re-established themselves as Scotland's second party at Westminster level, but have failed to do so at Holyrood level.  Five of the seven post-election polls, including tonight's, have had Labour ahead of the Tories at Westminster.  But five of the seven have also had the Tories either ahead of Labour, or level with Labour, on the Holyrood constituency vote.  The explanation for this phenomenon is probably a small degree of cross-voting - ie. a modest percentage of people who would vote SNP on the Holyrood constituency ballot, and either SNP or Green on the Holyrood list ballot, are backing Labour at Westminster.  That's proving just enough for the moment to keep Labour ahead of the Tories in Westminster terms.

Survation also asked the independence question...

Should Scotland be an independent country?

Yes 46% (n/c)
No 54% (n/c)

In a sense that corroborates YouGov's findings, which also suggested there has been no statistically significant changes in public opinion on independence.  The difference, however, is that YouGov continued to suggest the Yes vote is slightly lower than the 45% recorded in the 2014 referendum, whereas Survation are continuing to suggest it is slightly higher than 45%.  There seem to be 'house effects' at play here.  But whichever firm you choose to believe, the post-election bandwagon effect for No that was predicted by some has plainly failed to materialise.