“In most polls there are always about 5 percent of the people who ‘don’t know.’ What isn’t generally understood is that it’s the same people in every poll.” – George Carlin
(This blog post previously appeared in Common Weal’s weekly newsletter. Sign up for the newsletter here.)
Since 2017, I’ve been collecting and deeply diving into Scottish polling data around independence. The last full report was published a little under a year ago but with the re-launch of the independence campaign ahead of the independence referendum that the Scottish Government hopes to hold next October, now is a good time to revisit that study. While I keep an eye on all polling from all polling companies in Scotland I tend to restrict my deeper analysis to those published by Panelbase as they tend to break up their dataset into more varied subsets than others like YouGov and thus provide a richer story for those trying to find out not just how many people support independence but who they are. Since August last year, there have only been five Panelbase polls asking Scotland about independence, including the one just released this week, so there isn’t yet enough data to publish another full Demographics of Indy report. However, given that this latest poll is the first since the First Minster’s indyref announcement and it returned a majority for support for independence I felt it was worth giving a mini-update in this news column. Please read the full policy paper series – comprising the 2021, 2018 and 2017 editions – for my methodologies and all of the caveats involved in peering darkly through the lens of polling data.
Overall support for independence has been steadily climbing since the start of 2022 though it has not yet reached the previous peaks seen during 2020. Support for the Union was also steadily climbing throughout this period (in both cases, it appears that Undecided voters have been choosing sides as the debate intensifies) however that support dropped sharply in this latest poll. We are, of course, peering into changes that are within the statistical margins of error for polling so we will need more data to determine if this 1% lead for the Yes Movement is a statistical blip or part of a trend.
PREVIOUS INDEPENDENCE SUPPORT
Probably the strongest predictor of how you currently feel about Scottish Independence is how you voted in 2014 (assuming you did). More than 80% of people who voted Yes or No in 2014 would vote in the same direction today with 2014 Yes voters being particularly consistent in their views. The major reason that 2014 Yes voters no longer support independence seems to be the issue of Brexit, with signs that some of those who voted Yes in 2014 and Leave in 2016 having moved to No. Those who didn’t shift to No in 2016 are not likely to have shifted at any time since then. 2014 No voters, by contrast, appear to be more influenced by current events (perhaps the ongoing nature of Brexit rather than the mere fact of it) and much more willing to at least consider changing their mind. This works both ways for the independence campaign. It is easier for pro-independence activists to move No voters over to Yes than for pro-union activists to do the opposite, but one of the failures of the independence movement has been our lack of ability to “anchor” those newly won supporters – many of them appear to have swayed back to No.
Often lost in discussions about Yes and No voters are the Undecided voters. Those who didn’t vote in 2014 are looking at their options now and are much more evenly split on their current views than either 2014 Yes or No voters (as one might expect). However, 70% of this undecided group declared that they would vote Yes in the most recent Panelbase poll. Undecided, they might be but as we shall see in other categories, the “Undecided” voters look a lot more like soft-Yes voters than they do ardent advocates of the Union.
One important factor to note in the latest poll is the apparent hardening of views. Since the indyref announcement, Yes voters and Undecided have become more Yes, No voters a more No. While, as mentioned above, there is scope to pull in and anchor people who were undecided in 2014 and the recent evident suggests that they want to be convinced of the merits of independence, the movement may find it harder than it ever has been to pull someone across the divide from their previous No vote.
THE IMPACT OF BREXIT
Prior to the 2016 referendum on the EU, the overlap between Brexit attitudes and attitudes towards independence was notable but not as stark as it has since become. The four-way debate appears to have polarised such that there are now comparatively few Yes/Leave voters and No/Remain (or Rejoin) voters. This is further reflected in UK political attitudes where neither the Conservatives nor Labour hold a position of rejoining the EU and even the Lib Dems see it now only as a matter for the indefinite future. For those who wish for Scotland to rejoin the EU it is not surprising that they therefore see independence as the only available vehicle for this. Conversely, whilst Scotland rejoining the EU is not a guaranteed matter, the fact that the two largest interdependence political parties – SNP and the Greens – are themselves pro-EU means that anyone who sees Brexit as the defining political victory of the age appears to be less than likely to support a Scottish independence that would undo that effort. As with 2014 Independence supporters, those who were Undecided in the 2016 Brexit Referendum tend to “look” more like Remain voters than Leave voters – a fact that may be surprising considering that Remain was closer to the “status quo” vote than Yes was. If anything, the latest polls suggest that 2016 Undecided voters are even more pro-independence throughout 2022 than Remain voters have been with the latest poll placing independence support within these two groups at 66% and 52% respectively (compared to 30% independence support amongst 2016 Leave voters).
INDEPENDENCE AS A CLASS ISSUE
As stated in the main reports, one of the weaknesses of current polling is that we lack good data on income and class. The closest we have is infrequent breakdowns by income in some polls (not Panelbase and too infrequent to draw trends) and class measured by the infamous NRS Social Grade rating. However, it is best and only measure we reliably have to discuss the relation between social class and support for independence. In general, we have seen that the “Middle Classes” of ABC1 social grades are generally not supportive of independence whereas the “Working Classes” of C2DE have maintained consistent majority pro-independence since 2014. This high level of support for independence should be particularly noteworthy considering that the “E” group also includes pensioners and age itself is generally considered to be a strong correlator with support for the Union. This suggests that support for independence amongst younger working class voters will be even higher than 50%. Since the start of 2022 and especially since the announcement of the proposed independence referendum next year, both of these groups have polarised themselves in a manner similar to the breakdown by 2014 report with middle class voters increasing their support for the Union and working class voters increasing their support for independence.
GENDER AND AGE
Unlike many polling companies, Panelbase breaks its results down by both gender and age and allows us to see a particularly rich set of trends in independence sentiment.
The general picture since 2014 is that older people are less likely to support independence than younger people and that that age gap is generally widening rather than decreasing as might be expected if one subscribed to the hypothesis that support for the Union might “age out” of the Scottish population with time. Younger folk as a whole have steadily increased their support for independence since 2014 as have older females (whose support for independence has now almost rebounded from a shocking drop in independence support observed around the time of the Brexit referendum). Older males have steadily declined in support for independence and males in the middle age category have been particularly volatile and unpredictable. The biggest drops in support for independence since the indyref announcement were in 55+ females and 35-54 year old males whereas the largest gains were observed in 35-54 year old females and 16-34 year old males. Obviously this mixed picture suggests either a substantial amount of noise within the polling trends or that the trend and the causes of it are not yet readily apparent.
It should come as no surprise that voters have generally aligned their voting intentions along with their views on independence. SNP voters are overwhelmingly (though not unanimously) pro-independence whereas voters of the three main pro-Union parties are generally (though not unanimously) themselves pro-Union. (Panelbase only asks about voting sentiment towards the parties generally able to win seats in a UK General Election and doesn’t include an “Other” category). Since the indyref announcement, SNP voters may have increased their independence sentiment slightly though it remains within their long-term range. Support for independence amongst Labour and Lib Dem voters however has dropped markedly since last week’s announcement. Support for independence amongst Conservative voters remains at its historic average of about 5% which gives me the chance to once again remind people that despite this low level it is extremely likely that there are more pro-independence Conservative voters in Scotland than there are paying members of the Conservative party.
COUNTRY OF BIRTH
One of my biggest bugbears when it comes to Scottish independence polling is the lack of basic data on sentiment amongst “New Scots”. The best we have at all comes from Panelbase (which is one of the reasons I study their polls more closely than those from other companies) and even that is woefully insufficient. Results are broken down by country of birth only into three categories. Scotland, England and “Other”. This means that the poll combines results for EU and non-EU immigrants with immigrants to Scotland from Wales and Northern Ireland. I understand the reasons for not breaking sub-samples down into much smaller categories (see the 2021 Demographics of Independence paper for details of that methodology) but I really would plead with polling orgs to at least provide a column showing Country of Birth as Scotland, rUK and Other (and for efforts to be made to break “Other” into EU and Non-EU).
Such that it is though, “Other” voters are now by far and away the most pro-independence demographic in Scotland, exceeding even those born in Scotland by a substantial margin and have hit their highest level of support for independence ever. Voters born in England have seen a similar boost in support for independence, also approaching record levels ever. The boost amongst voters born in Scotland has been much more muted however though the worrying drop in support observed throughout late 2020 and into 2021 appears to have arrested in 2022 and at least partially recovered.
One factor that comes through strongly in polling and will have a major impact on overall results is in the likelihood that respondents will vote at all. This is a tricky area in polling as the percentage of people who will tell a pollster that they intend to vote is almost always higher than the percentage of people who actually turn out to vote (due to a combination of the natural inclination to not tell a pollster that you’re not going to do your civic duty and the fact that folk who are particularly disengaged or disgruntled with politics are more likely to not respond to a pollster at all). However, what we do see in the latest of these Panelbase results is that older, middle class and No-leaning people are substantially more likely to vote than younger, working class and Yes-leaning people. We also see that people born outwith either Scotland or England are much less likely to say that they will vote in a future independence referendum. It obviously does the Yes Movement little good if we convince people of our cause but cannot convince them to turn out to vote. This was a particularly acute problem in the 2014 referendum where organisations like the Radical Independence Campaign did absolutely sterling work to engage with and turn out working class voters in cities like Glasgow and Dundee which saw their voter turnout on the day reach historic record levels. However, it is also true that despite this effort, those historicity high turnouts still resulted in Glasgow and Dundee seeing some of the lowest levels of voter turnout across the whole independence referendum.
As we campaign across the next 16 months the independence movement must look at how its arguments attract voters across the divide (a challenging task given the hardening of views and the possibility that policies aimed at those voters may well alienate others already otherwise convinced of independence) but more importantly we should look at how we can increase turnout amongst those who already agree with us but may, for any reason or none, not put pencil to paper on the day.
WHO SHALL VOTE FOR US?
One difference between 2014 and now is that there is presently great uncertainty about how Scotland will democratically demonstrate its support for independence and the manner in which we do so – either via a referendum or a UK General Election plebiscite – will have a substantial impact on who gets to cast their votes. It is extremely unlikely that the UK Government will change the UK General Election franchise to match the franchise currently in use in Scottish elections and referendums. This means that if the option of a sanctioned or unsanctioned referendum are not possible and a plebiscite is the only viable route then this will result in the disenfranchisement of 16 and 17 year olds and most EU and non-EU immigrants to Scotland. As we have seen above, these groups are some of the most pro-independence demographics in Scotland at the moment – based on known demographics and some approximate calculations based on the most recent polling, disenfranchising these groups could result in a net loss of around 120,000 Yes votes. This immediately makes a plebiscite a less attractive route to independence from the point of view of existing voting sentiment and it also means that until and unless the uncertainty is resolved, it becomes harder even to campaign towards voters who don’t yet know if they’ll even be allowed to vote (not that this prevents such folk from campaigning – I could well see disenfranchised voters campaigning for others to “vote for them” if they cannot).
This brings up back to voting turnout though. The independence campaign must consider how it can reach people who are already minded to vote Yes, we know will be able to vote Yes regardless of a referendum or a plebiscite but are likely to not vote or not even be registered to vote unless we do something to correct that. What RIC did in 2014 was wonderful and was inspiring but it was simultaneously not enough. That campaign to turn out the disenchanted, those whom the Union has done more to break down than anyone else and those with the least to lose and the most to gain from independence. Between now and our next chance at an independence vote, we need to repeat that work and we need to exceed it. Do this, and independence truly will be within our grasp.