Every time there's a poll on independence that shows a No lead, there's always a minor chorus of people asking me "did they interview 16 and 17 year olds?" And it's been getting to the point where I just roll my eyes to the heavens, because the answer is almost always "yes" - it's become completely standard to include over-16s in Scottish indy polls. But, I'm afraid, the now-notorious Deltapoll survey for The Sun is a shocking exception to that general rule. Marcia pointed me in the direction of the newly-published datasets this morning, and if they're accurate they show only over-18s were polled. There's no way of knowing whether that made any difference to the headline results, but it could conceivably have cut the Yes vote by 1% - in other words if 16 and 17 year olds had been there (as they certainly should have been), it could have been a 52-48 lead for No, rather than 53-47.
I suspect the reason for this breach of good practice is simply that Deltapoll are not experienced in running Scottish-only polls - it may not have occurred to them that 16 and 17 year olds needed to be there, or they may not have had any 16 or 17 year olds on their panel. The explanation for the small sample size may also be that they don't have enough Scots on the panel - it seems unlikely The Sun didn't have the cash to pay for a full-scale sample of 1000.
There are a few other points of interest. The poll shows that only 47% of the sample (before Don't Knows are excluded) are opposed to an independence referendum. This will come as a massive surprise to anyone who read the Sun's write-up of the poll, which claimed that 55% were opposed to a referendum. They can only have reached that figure by excluding Don't Knows - but if that's what they did, why did they quote a Yes figure of 42% which they arrived at by not excluding Don't Knows? It's just the most appallingly deceitful and cynical article you'll ever see - it leaves Don't Knows in when it's convenient, excludes them when it's not, and never even bothers to mention that different numbers are being calculated in a completely different way. (And frankly, I suspect the 55% figure must be yet another outright lie. The figures with Don't Knows excluded aren't available on the data tables, but I can't see any arithmetical way in which 55% is even possible. The real figure is probably around 51%.)
There were only ten Green voting respondents in the sample, and they had to be massively upweighted to count as 36 people. That's magnified a peculiar (and almost certainly inaccurate) pattern - 71% of Green voters are counted as Yes supporters, zero as No supporters, and an unusually high 29% as Don't Knows.
There's also a finding in the poll that shows the Queen's death has made 17% of people more likely to vote for an independent Scotland, compared to only 12% of people who are less likely to vote for independence. Mysteriously, this result is totally absent from the Sun's report, and I'm sure that's got nothing whatever to do with the fact that it drives a coach and horses through the "independence support plummets due to Queen's death" angle they wished to take.
Incidentally, the Sun article has now been quietly modified. The word "plummets" has been replaced by "falls", and the claim of a "seven-point drop" in Yes support has been replaced by "four-point drop". (In reality, it's a mere two-point drop on the headline numbers, and it's a comparison they shouldn't even be making because the polls were conducted by different firms, but they can justify the claim of four points by leaving Don't Knows in.) This almost certainly means they knew they were on a sticky wicket with the press regulator IPSO. If anyone is proceeding with an IPSO complaint, my strong advice would be to not accept this furtive change as sufficient and to push instead for a clear correction and apology. And please note that the Express article about the poll has not yet been changed to remove the false claim about a seven-point drop in Yes support.
You might like to know there's an article in today's edition of The National, which quotes me at length about The Sun's lies - you can read it HERE.
* * *
Scot Goes Pop General Fundraiser
Scot Goes Pop Polling Fundraiser
If you prefer another method, such as Paypal or bank transfer, please message me for details using the contact email address which can be found in the sidebar of the blog (desktop version only), or on my Twitter profile.