Sensational Social Attitudes Survey suggests that independence support has soared by almost TWENTY percentage points since the 2014 indyref - does this suggest that conventional polling has been underestimating Yes support for years?

I've finally had a chance to catch up with the independence numbers from the Social Attitudes Survey that caused so much excitement a couple of days ago.  What strikes me most is that the survey seems to exist in a slightly different universe from the more conventional polling we're used to seeing.  We thought we knew that 2020 was (by far) the high watermark for independence support to date, and that 2021 was a step backwards, but this new survey suggests that support actually reached a fresh peak in 2021 - albeit there was no Social Attitudes Survey in 2020 itself for obvious reasons, so it's possible an even bigger number was missed during the gap.  Nevertheless, an outright majority for independence on a multi-option question is a startling finding given the pattern of polls in 2021.  It could perhaps be more easily explained if the fieldwork had been conducted at the start of the year, because the unbroken sequence of Yes majority polls carried on into January 2021 (as I remember extremely well, because I commissioned a Survation poll myself that month which showed Yes on 51%).  But in fact the Social Attitudes Survey was conducted in the autumn, and was thus contemporaneous with a series of eight polls in a row which showed a No lead, albeit mostly a fairly modest No lead.

2021 Scottish Social Attitudes Survey constitutional preferences (with changes from 2019):

Independence: 52% (+1)

Devolution: 38% (+2)

No Scottish Parliament: 8% (+1)

Don't Know: 1% (-4)

The main difference of format between this survey and conventional polls is that respondents weren't asked to give a binary Yes/No answer to independence, but instead to choose between multiple options, including more than one option for expressing support for the union.  You would normally expect such a format to produce less support for independence than the conventional polls, and yet for some reason it's produced more.  

But, actually, I would tend to suspect that if a binary question had been asked in the survey, it might have resulted in an even bigger Yes vote.  I think there's something going on here that isn't about the question format - I think it might be about the data collection method.  The Social Attitudes Survey has traditionally been conducted face-to-face, which makes it very different from the online polls we're used to seeing.  If the 2021 survey had been face-to-face as usual, that would be a very obvious potential explanation for the divergence with conventional polls, and might call into question whether online polls have been systemically underestimating the Yes vote for some time - especially bearing in mind that the only telephone poll conducted in autumn 2021 (by Ipsos-Mori) also contradicted the online polls by showing a very handsome Yes lead.

But just to muddy the waters, the pandemic meant that the Social Attitudes Survey wasn't conducted face-to-face last year - the What Scotland Thinks website suggests it was a hybrid telephone/online panel survey.  Nevertheless, I do still wonder if the telephone element, and perhaps a different approach to online fieldwork (particularly random participant selection), may have been factors in producing such a strikingly different result from conventional polls.

The other sense in which the Social Attitudes Survey seems to exist in a different universe is the longer term trend.  It suggests that support for independence has more than doubled since 2012, has increased by almost twenty percentage points since the indyref year of 2014, and has increased by thirteen points even since the SNP's golden year of 2015.  There's no sign in conventional polling of quite such a dramatic transformation, and Stuart Campbell must be looking on in horror, because it's a trend that's utterly impossible to reconcile with his repeated claims of flatlining Yes support since 2014.  The question format probably is the explanation here, because it looks as if a lot of people who said they supported devolution in past surveys nevertheless voted for full independence when presented with a binary choice in a referendum.  But their increasing readiness to actually identify as independence supporters rather than devolution supporters suggests that their position has hardened considerably as the years have progressed.

And just a final thought.  Renowned Liberal Democrat "fast bowler" Alex Cole-Hamilton tried to mock The National for reporting a survey with fieldwork that was a year old.  "Whatever this is, it's not journalism" he sneered.  But hang on.  Isn't it his own side that claims that the state of public opinion eight years ago is more important than the state of public opinion now?  Even though in 2014 nobody knew Brexit was coming, and even though some people who are now eligible to vote were only eight years old back then?

*  *  *

SCOT GOES POP NEEDS YOUR HELP TO KEEP GOING, AND ALSO TO COMMISSION MORE OPINION POLLS ABOUT INDEPENDENCE

If you've donated already, a million thanks!  And if you'd like to make a donation now, here are the various options...

Direct payments via Paypal - my Paypal email address is:   jkellysta@yahoo.co.uk

(Paypal payments are the best method because they're direct and eliminate all fees as long as you choose the "paying a friend" option.  However, please take great care to spell the above email address correctly.  Also, if you wish you can add a note saying something like "for the fundraiser", but rest assured it'll be obvious what the payment is for anyway.)

Scot Goes Pop General Fundraiser 

Scot Goes Pop Polling Fundraiser 

If you prefer a 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.