Clueless Sun newspaper left red-faced after being caught wrongly telling readers that support for independence has "dropped"

A couple of nights ago, someone left a comment on this blog asking me to write about the numbers from the new Progress Scotland poll on support for independence.  I wasn't sure what he/she meant, because none of the reporting I'd seen on that poll mentioned any direct question about independence.  I checked the Progress Scotland website and there was no clue there (and frustratingly there was no link to any datasets).  The mystery deepened when The Sun started reporting that the poll supposedly showed Yes support had "dropped below 40%".  At that point my propaganda detector started buzzing very loudly and I was fairly convinced that things were not quite as they seemed, but obviously it was hard to comment until I actually saw the relevant part of the poll.  I was finally pointed in the right direction by Denise last night - the datasets had been on the Survation website all along.  (In fairness to myself I had already looked there, but the archive section is a bit confusing.)

And the verdict?  You probably won't faint with amazement upon learning that The Sun have been telling you porkies.  The lie is not the suggestion that the poll shows support for independence at below 40% - that can just about be justified at a stretch, albeit it's somewhat misleading.  No, the lie is the claim that Yes support has "dropped" in the poll, and that there are serious implications for Nicola Sturgeon, indyref2, etc, etc.  This implies that there are comparable polls from the recent past which showed much higher support for independence - in fact, there haven't been any such polls.  The Progress Scotland poll asked respondents about their views on independence in a novel way, and also asked them to give their answers in a novel way by rating their support or opposition to independence on a scale of 0 to 10.  There is nothing to compare the results to - which means that support for independence may have risen, or it may have fallen, or it may have stayed exactly the same.  The poll gives us absolutely no clues to trends whatsoever.  But what we do know is that a Survation poll a few weeks ago, which asked the independence question in the conventional Yes/No manner, showed Yes support holding steady in the mid-40s.  It's therefore not unreasonable to work from the assumption that nothing much has changed, at least until we see the slightest scrap of evidence to the contrary.  The likelihood is that a conventional independence poll conducted now would show a no change position, and that a poll with a 0-10 format conducted a few weeks or months ago would have produced much the same results as the Progress Scotland poll.

Now, let's be fair - it's theoretically possible that there has been a sudden and inexplicable drop in the Yes vote in the few weeks since the Survation poll.  But there are two possible explanations for the Progress Scotland numbers being so different from the norm - a) a real drop, and b) the use of a radically different question and answer format.  As b) is so obviously the simplest and most natural explanation, the onus must surely be on The Sun to substantiate their outlandish claims.  My own question would be whether they were deliberately pulling the wool over their readers' eyes, or whether they really were totally oblivious to the fact that they were comparing apples with oranges. I've learned from experience never to exclude the possibility of utter journalistic cluelessness.

So why might the different poll format have such a dramatic effect on people's answers?  It may be the reference to "staying part of the UK" - there are plenty of precedents for polls that use that sort of language producing much more No-friendly results.  Or it could be that soft voters in the middle, or undecided voters, react differently when they're not presented with a binary choice.  Either way, we shouldn't jump to the conclusion that the poll is picking up underlying resistance to independence in a way that conventional polls have not.  If the Yes/No polls had been that far out in 2014, the eventual Yes vote would have been nowhere near as high as 45%.  It's true that the polls did overestimate Yes, but not by much, and it's possible that some or all of that discrepancy can be very simply explained by differential turnout.

Oh, and as for Pamela Nash and Scotland in Union taunting Progress Scotland for selectively withholding the least favourable results from the poll, the words pot, kettle and black spring to mind.

*  *  *

I was contacted on Twitter last night for my views on whether SNP voters in the forthcoming Leith Walk by-election should use their lower preferences.  Answer: yes!  Your lower preferences will not be taken into account unless the SNP candidate is eliminated before the final two, but in the unlikely event that it pans out that way, I'm sure most of us would have a view on whether we'd rather have a Green councillor or a unionist councillor.  (The Greens are strong in the ward.)  The rule of thumb is always the same with the STV voting system - use as many preferences as you feel able to, because you'll never be doing any harm, and you'll sometimes be doing a lot of good.