How unionist MSPs and the Express made idiots of themselves last night by celebrating FICTIONAL poll numbers on independence: what we know so far

I think I've finally pieced together what happened last night. It looks like the Panelbase poll that unionist MSPs were "celebrating" was the one commissioned by the Alba Party which I reported on a few days ago. When that poll was published, these were the results -

Should Scotland be an independent country? (Panelbase, 7th-11th October 2022)

Yes 49% (-)

No 51% (-)

The dashes in brackets indicate no change at all from the previous Panelbase poll in August, which was also Yes 49%, No 51%.  So nothing for unionist politicians to celebrate there - quite the reverse in fact, because the numbers show the BBC propaganda following the death of the Queen failed to make even the slightest dent in support for independence.

However, another Panelbase poll commissioned by Believe in Scotland was published after the Alba poll, even though it had actually been conducted a few days earlier.  It showed Yes 50%, No 50% - a considerable psychological breakthrough for Yes, because it's a relatively rare example in this calendar year of a poll that doesn't have No with a small lead.  But the downside is that it means we have to retrospectively update the percentage changes on the Alba-commissioned poll, because that one was conducted later.  The correct figures on the Alba poll are now...

Yes 49% (-1)

No 51% (+1)

When you have three successive polls from the same firm showing nothing more than a tiny fluctuation up and back down again of just one percentage point, that's practically the dictionary definition of 'margin of error noise' - meaning nothing has changed and the fluctuation is caused simply by normal sampling variation.  So, again, nothing for unionist politicians to celebrate.

But then some random bloke on Twitter called 'British Electoral Politics' decided to misreport the Alba poll as showing Yes 48% (-2), No 52% (+2).  When challenged as to where these inaccurate numbers had come from, he brazenly acknowledged that he hadn't actually seen the results of the poll with Don't Knows excluded and had made them up himself based on a crude recalculation of the figures prior to the exclusion of Don't Knows, which were the only ones he had seen.  He had made no disclaimer when publishing the fake numbers, which led to the cartoonish spectacle of the Express and heaven only knows how many unionist politicians (including Stephen Kerr and "fast bowler" Alex Cole-Hamilton) proclaiming the end of the independence cause, based solely on a work of fiction.

I've seen the raw numbers from the Alba poll and they show that 407 respondents (48.6%) said they would vote Yes, and 431 (51.4%) said they would vote No.  The numbers have been correctly rounded in the published results (I believe they were originally published as long ago as Friday in The National) to Yes 49%, No 51%.  The Express appear therefore to have published yet another flat-out lie about Scottish independence polling.  I'll wait until the end of the day to see if there are any further twists in the tale, but if not, and if the Express don't correct their error, we might have to think about complaining to the press regulator again.