We wrap up the first batch of our 2021 election posts with a deep-dive on the relative strength of support for Scottish independence versus staying in the union – and our findings might come as a bit of a surprise. Stay tuned for more essential insights on the 2021 election and Scottish public opinion. Header image © User: Kyoshi Masamune / Flickr / CC BY-NC 2.0
It didn’t take the Holyrood election to tell us that Scotland is split down the middle. We already know from the opinion polls that support for independence is, like the SNP as the votes were counted, trembling on the brink of a majority.
Where surveys like the ESRC-funded Scottish Election Study (SES) add value is in looking beneath the surface of election results and poll ratings. Clearly, the Yes and No camps are roughly much equal in size, but how about in terms of conviction and commitment? Who is winning in the race to that movable finish line of #indyref2?
The first thing to say – surprising no one who has been living in Scotland since 2014 – is that conviction levels on both sides are very high. SES 2021 respondents were asked after the May election about the certainty of their voting intention in a second referendum and given a scale with no fewer than 21 points to choose from. Two out of three respondents chose one of the extreme points – ‘would definitely vote Yes’ (29%) or ‘definitely vote No’ (37%) – and most of the others were close to these poles. This already shows an advantage for the pro-union side – of which more later – but it also demonstrates that both sides have a very solid base of support.
Does that mean that any future referendum campaign will be held simply in pursuit – bordering on persecution – of the fewer than 10% of people who are in the undecided zone? Not necessarily. For one thing, we know from a similar question in 2014 – asked less than three months before polling – that even those towards the extremes on these scales reserve the right to change their minds. The shape of the 2014 and 2021 curves is very similar – and both risk overstating the entrenchment of opinion.
For another, we know that minds have been changing since 2014 – and more than is implied by the fairly small and slow movements in the polls. While those polls indicate an appreciable shift to Yes, 2021 SES data reveal that that is due more to generational replacement: new entrants into the electorate are more likely to be Yes voters while those who pass away are likelier to have supported No. The changing of minds has moved equal shares of voters in either direction: 16.4% of 2014 Yes voters now support No, while 16.6% have gone in the other direction. These individuals don’t quite balance each other out on net – it benefits the Yes side slightly because there were more No voters to begin with – but it is remarkable how similar these shares are.
Will those switchers stay put? Again, the No side appears to have an advantage in terms of how zealous its new converts are. On the 21-point scale discussed above, 58% of Yes to No switchers consider themselves certain to vote for their new preference versus 40% of No to Yes converts. While a large share of new Yes supporters are pretty confident they’ll stay in that camp – a further 35% selected values between –5 and –9 on the scale – the pro-independence camp clearly have more work to do in retaining incoming supporters.
And what about potential future switching? There are further signs of asymmetry there. One we have already seen in the first graph above. The No side not only has more respondents on its side of the scale (49% compared to 44% for Yes) but it has a particular advantage (37% to 29%) among those reporting that they would ‘definitely’ vote that way. The No advantage persists when we isolate respondents who haven’t changed their mind since 2014, with 43% being definitely No vs. 32% for definitely Yes.
Parallel results come from a set of questions asking how far being Yes or No has become part of voters’ identities. A recent study shows that, across Britain, Remain and Leave became a more strongly-felt set of identities than voters’ allegiances to the political parties. In Scotland, the same is true but even more so when it comes to Yes and No. However, the graph below shows that No has a clear advantage when it comes to strong identities.
It might be thought that this simply reflects No’s advantage in the survey sample overall. But the asymmetry goes deeper than that. Among those reporting a No vote intention in a second referendum, 70% felt very strongly identified with that side. This was true of only 61% of those intending to vote Yes.
Why might there be this asymmetry in conviction across the two sides? We end with two suggested reasons, one based on the substance of policy, one based on voter psychology.
The policy point starts with the idea that, in between the status quo and some notion of ‘full-blown’ independence, there are some intermediate stages that are seen as belonging on the No side. From No all the way to Yes, then, might be a bigger distance than from Yes to a version of No that nonetheless offers change from an unpopular status quo.
This rather abstract point is illustrated by the table below, which uses a question from the 2014 Scottish Referendum Study that asked those on each side of the debate about various things that “might tempt them” to change their mind. Such hypothetical questions have their flaws and the absolute percentages in the table should be treated with caution. But those flaws should not distort the key comparison which reveals that Yes minds look easier to change, especially by steps in the direction of further devolution.
|Devolution of full taxation powers||14|
|Relocation of Trident outside Scottish waters||10|
|Control over all revenues from oil in Scottish waters||18|
|If an independent Scotland could not use the pound||3|
|None of these||71|
|If no plans for further devolution||9|
|If looked likely that the Conservatives win 2015 UK election||8|
|If polls indicated UK would vote to leave the EU||7|
|If UK agreed to currency union with independent Scotland||6|
|Nothing would change mind||77|
The second possible reason why No voters are more certain is simply that they are older – and age breeds conviction (whether this is the confidence of experience or the closing of minds is a moot point for present purposes.) This much is confirmed by the graph below. It shows, separately for the Yes and the No sides, how the proportion of people at the ‘Definitely’ end of the vote certainty scale varies by age category. On both sides, conviction increases with age – at least until around 70. The problems for the Yes side are twofold, then: i) its vote is drawn disproportionately from the more equivocal age range; ii) even when it does attract older voters, these are not the most convinced.
The polls place Yes and No more or less neck and neck. The SES data in this blog post point to a small but appreciable advantage for the No side in terms of the conviction underlying those vote intentions. This may come as a surprise given that one story of the referendum campaign was a major asymmetry in the other direction: there seemed to be stronger feelings, more passion and certainly more participation among Yes supporters.
The SES data capture that, too: for every respondent who reported going on an anti-independence demonstration, almost five reported attending a pro-independence event. This is not unusual in a case where the status quo confronts change: it is the latter that tends to motivate and mobilise. Nevertheless, a polarising event like the 2014 referendum can entrench much of even a relatively silent majority. There are just about enough voters in the middle ground for either side to win a second independence referendum, but the Yes campaign’s task looks harder based on these data – especially the contest happens sooner than later.