Last month, we published two posts on Scotland’s “Four Tribes”, the groups the country’s electorate can be divided into on the basis of public preferences on Scottish independence and European Union membership. In the first, we compared the groups in size and broke down how they voted in the 2019 UK general election. In the second, we walked through their demographic characteristics using an interactive visual story. This time around, we’re wrapping up the series by examining their identities and political beliefs. How do the Tribes identify, and what does that mean for their views?
To recap, the Four Tribes are No/Leave, No/Remain, Yes/Leave and Yes/Remain. In our first post on the topic, we showed that the Remain Tribes both outnumber the Leave Tribes, and that the Yes/Leave grouping is the smallest. The No/Leave and Yes/Remain groups each overwhelmingly vote for the Conservatives and the SNP respectively. The biggest group, No/Remain, was very evenly split between the four main parties in 2019. Meanwhile, around half of the the Yes/Leave grouping – the only one of the four without an obvious party political tribune – clearly prioritised Scottish independence over Brexit by voting for the SNP. As is the case in the rest of the UK, the Remain Tribes are significantly younger than the Leave Tribes and include a greater share of university graduates.
Readers may be surprised to find that sociodemographic differences are significantly less pronounced across the Yes-No divide than the Leave-Remain one. While there are of course some important differences between social groups when it comes to independence – age being the most obvious – it is remarkable how little these characteristics explain about Scottish constitutional preferences. Compared to Brexit, this divide is much more to do with values and identities, and we can again explore these using data from the last Scottish Election Study, carried out before the aforementioned 2019 general election.
The first graph below, a version of which was also shown at the end of our second post on this topic, shows histograms of how each Tribe is distributed along a 21-point “Relative Territorial Identity” scale (RTI). This measure was first introduced by authors including SES team members Prof. Ailsa Henderson and Dr. Jac Larner in a recent peer-reviewed publication. It is calculated by subtracting respondents’ Scottish identity from their British identity, both of which are recorded on 0-10 scales. This produces a 21-point “relative” scale from -10 to +10. The higher the score, the more “Scottish” a respondent feels compared to British. A score of 0 means the person feels these identities with equal intensity, whereas a -10 means they feel completely British but not at all Scottish and a 10 means the opposite. Meanwhile, a 3 on the scale, for example, would mean that the respondent feels 3 points more Scottish than British. That difference could be between, say, 7 and 10, or 3 and 6; the absolute values don’t matter so much, and this is why we refer to the scale as “relative”. While we obviously sacrifice some information about the total strength of respondents’ identities using this scale, its big advantage is that it’s more fine-grained than the classic “Moreno question”, which asks respondents to directly trade-off their Scottish and British identities on a five-point Likert scale (the most common responses to which are usually “Equally Scottish than British” and “More Scottish than British”).
In a histogram, a higher bar on the vertical axis means more respondents land on the relevant value represented on the horizontal axis. The horizontal axis in the figure below shows each value of the 21-point RTI scale. These graphs show that a substantial plurality (the biggest share if there is no majority) of anti-independence voters, around one third within each No Tribe, land on a 0, meaning they feel equally Scottish and British. A visibly higher share of No/Remain voters than No/Leave voters are found to the right of this midpoint, meaning they identify as more Scottish than British. There is also a bigger spike at the far left of the chart for the No/Leave group, with around 15% of these voters describing themselves as completely British but not at all Scottish. In other words, No/Leave voters are the most ardent British identifiers. The No/Remain histogram is the closest of all to a “normal distribution” (also known as a bell curve), with respondents clustering in the middle and generally tapering off at the far ends, which suggests that this group is the most moderate on matters of national identity. It is worth noting that, for both No groups, the vast majority of those at the very bottom of the scale were born in elsewhere in the UK, so this is not primarily to do with Scots-born voters eschewing any and all Scottish identity.
Unsurprisingly, the Yes groups are more likely to prioritise Scottishness over Britishness. This is especially the case for the Yes/Remain group, more than 20% of whom rate themselves completely Scottish and not at all British. In this case, then, we do observe a reasonable share of respondents consciously rejecting an identity they could feasibly lay claim to. Furthermore, very few people in this Yes/Remain group identify as more British than Scottish, and 0 is only the fourth most common value. Turning to the Yes/Leave Tribe, while its RTI distribution is more similar to the Yes/Remain one than those of the No Tribes, there is more variability in this group. As many Yes/Leave voters score a 0 as do a 10, and the share identifying as more British than Scottish is less trivial.
We can examine these distributions in more detail using boxplots, which visualise summary statistics. The vertical bar in each box below shows the median, the middle value in the distribution. The left and right sides of each box are the lower and upper quartiles, meaning the 25th percentile and 75th percentile of the distribution (i.e. the box contains 50% of respondents in each group). The mean for each group, the simple average, is shown using a diamond with a white border.
These plots show even more clearly that the lean to each identity is not symmetrical across Yes and No Tribes. In other words, the Yes Tribes identify as much more Scottish than the No Tribes do British. The mean and median for Yes/Remain are both 5, meaning that the average member of this Tribe identifies as 5 points more Scottish than British. The mean for the mirror image No/Leave group, on the other hand, is only around 2, while the median (and upper quartile) is 0. The mean being so distant from the median in this case is due to the strong negative skew of the No/Leave distribution; while most of these respondents are equally Scottish and British or slightly more British than Scottish, the simple average is pulled down by a substantial minority at the very bottom of the scale. Once again, the No/Remain group is the most balanced. It is also quite concentrated around the midpoint, with this Tribe having by far the smallest interquartile range. The difference between the Yes Tribes, meanwhile, is more obvious when visualised in this way.
Now that we’ve established how the Four Tribes identify, let’s take a look at what they think. The next graph shows how each group is distributed on a 0-10 left-right scale on which respondents were asked to place themselves. 0 means the furthest left and 10 means the furthest right. The demographic and partisan profiles of each Tribe are strongly suggestive as to their ideological leanings, and these expectations are reflected in the histograms below. The No/Leave and Yes/Remain groups lean clearly right and left respectively, the latter more strongly. Paralleling the RTI scale, the Yes/Remain grouping is the only one for which the scale midpoint is not also the mode. This group exhibits very consistent tendencies in its ideology, identity and voting behaviour, making it probably the most cohesive of the Four Tribes.
The No/Remain group is again relatively normally distributed, has a slight left lean and demonstrates something of a bias toward moderation, with sharp drops in the proportion of respondents placing themselves at a 2 to compared to a 1 and an 8 compared to a 9. This creates an interesting contrast between this group and its direct opposite, Yes/Leave, which has a similar centrist lean but is less evenly distributed. Nearly one third of these voters place themselves in the dead centre at 5, but there are also more voters at the absolute extremes; around 10% overall. It is worth noting that a similar proportion of No/Leave voters also selected 5 on the scale. Selecting the midpoint on this ideological scale may reflect apathy or a lack of political interest or knowledge, so it would make sense that a high proportion of Leave voters would select this option given known associations between Leave voting, education levels and “left behind” areas.
Considering both its members’ identities and left-right profile, the Yes/Leave Tribe is probably the most idiosyncratic and least coherent of the four, and this is likely related to its lack of political representation. “Elite cues”, the signals politicians send about their favoured issues and policy positions, are known to be important for shaping public opinion both as a general phenomenon and in Scotland specifically. The direction of causality in this regard – whether the Tribe’s lack of partisan attachments explains its scattergun ideological profile or vice versa – is not, however, immediately obvious. Since this is the smallest group, and it has no obvious ideological profile, it’s unclear if a pro-independence, anti-EU electoral ticket – something the disgraced former First Minister Alex Salmond’s new party Alba is flirting with – would be able to gain much of a foothold by appealing directly to this group.
Finally, which underlying attitudes or beliefs might be associated with these differences between the Four Tribes? In 2019, Scottish Election Study respondents were asked the extent to which they agreed with the statement “In Scotland our values make us different from the rest of the UK”. When we exclude “Don’t Know”s and amalgamate the “Strongly Agree”/”Tend to Agree” and “Strongly Disagree”/”Tend to Disagree” questions, we find starkly divergent views as shown below.
There is near-unanimity on this question among Yes/Remain voters, with more than 80% agreeing to some degree that Scottish values make the country distinct from its neighbours (more realistically, one particular neighbour with which the country shares a land border). For decades, the cause of Scottish independence has been closely associated with opposition to “Tory rule” at Westminster and the associated democratic deficit that caused so many ructions in the Thatcher era. The SNP in government have deliberately couched their pro-independence advocacy in a nationalism which is civic and inclusive rather than ethnic and exclusive. This, coupled with the largely fictional idea that Scotland’s population harbour significantly more socially democratic views than their English counterparts – “a mythological fusion of ideology with genealogy” – have led to an unusually strong relationship between left wing politics and nationalism. Many left-wingers in Scotland support independence for instrumental reasons, and this view is predicated on the assumption of Scottish political exceptionalism rather than ethnocentrism.
The rest of the Tribes fall into a, by now, predictable pattern. Yes/Leave voters lean the same way as their Yes/Remain counterparts, but agreement is closer to 60% than 80%. The No/Remain Tribe are once again remarkably evenly divided, with the Agree and Disagree categories each containing around 35% of these respondents and the middle option containing the remainder. The No/Leave Tribe is by far the most sceptical of this idea, with just under 15% of these respondents agreeing. Members of this Tribe tend to believe Scottish and British values are one and the same.
Thinking about how this intersects with the territorial labels discussed above, it seems Yes voters (especially Yes/Remainers) are motivated by a perceived conflict between Scottish and British identities, while Scots-born No/Leave voters tend not to perceive this. If it ain’t broke, why fix it? This discussion highlights why feelings on the matter of Scottish independence are so strong. They tap into, and in some cases threaten, the deeply-held identities and political values of a great many people.
To help better understand the social ripple-effects of the political conflict driven by these identities, vote choices and labels, the 2021 Scottish Election Study will investigate so-called “affective polarisation”, which means personal animosity between different opinion- or identity-based groupings. Researchers have already shown that, in the UK as a whole, Leave and Remain identities are as powerful as party identification in driving such social prejudices. In Scotland, as our blog serious has shown, this is complicated by the existence of a second, and even more divisive, layer of existential constitutional identity politics. Whatever course Scotland takes in the coming years, understanding how this manifests will be crucial to understanding why.