The decision by a narrow majority of British voters to leave the European Union (EU) in 2016 was a political earthquake that few had seen coming. It produced new political divisions, not only between the United Kingdom and the rest of Europe but also within the United Kingdom. In particular, the referendum campaign and the outcome generated two new political identities: “Leavers” and “Remainers.” These Brexit identities crosscut partisan identities and voters formed deep emotional attachments to them (Curtice Reference Curtice2018; Evans and Schaffner Reference Evans and Schaffner2019). Moreover, this Brexit divide led to affective polarization in the form of out-group animosity and discrimination (Hobolt, Leeper, and Tilley Reference Hobolt, Leeper and Tilley2021). It also shaped perceptions of the economy (Sorace and Hobolt Reference Sorace and Hobolt2021), attitudes toward immigration (Pickup et al. Reference Pickup, de Rooij, van der Linden and Goodwin2021), vote choices (Hobolt and Rodon Reference Hobolt and Rodon2020), and losers’ consent (Schaffner Reference Schaffner2021; Tilley and Hobolt Reference Tilley and Hobolt2023a). Brexit identities have been shown to be salient and politically consequential. Yet, we know much less about whether these new identities are rooted in policy norms that go beyond preferences about the desirability of leaving the EU. In this article, we thus explore the nature of Brexit identities and how they relate to policy norms.
We know from the literature on partisanship that political identities have important implications for people’s behavior (Campbell et al. Reference Campbell, Converse, Miller and Stokes1960; Green, Palmquist, and Schickler Reference Green, Palmquist and Schickler2004). This is due largely to the way in which social identities (e.g., partisanship) define membership of in-groups and out-groups (Tajfel Reference Tajfel and Tajfel1978). These group memberships also include a set of social norms: shared ideas and rules of behavior that people expect other group members to follow (Groenendyk, Kimbrough, and Pickup Reference Groenendyk, Kimbrough and Pickup2022; Suhay Reference Suhay2015). Indeed, a crucial element of partisanship is that these shared policy norms enable in-groups to agree on policy aims and provide cues to people regarding the “correct” policy choice for their side. For example, a Labour partisan identity in Britain is associated with support for redistribution and trade unions. But what about Leavers and Remainers? Other than the obvious agreement within each group about whether leaving the EU is a good thing, we do not know whether Brexit identities are similarly rooted in shared policy norms. In this article, we therefore examine the extent to which policy norms are important for people with a Brexit identity in a way that is distinct from their partisan identity. Specifically, we examine how policy preferences correlate with Brexit identity (“What do people think?”); how perceptions of policy norms vary by Brexit identity (“What do people think other people think?”); and, using a survey-embedded experiment, whether people are willing to align their own preferences with in-group preferences (“Do people want to think like their in-group?”).
We demonstrate that Brexit identities are associated with specific policy preferences and that these group norms are relatively well known to the public. We also show that providing policy cues about Brexit group-identity norms appears to be effective in shaping opinion. Overall, we contribute to the growing literature around nonpartisan political identities and their importance in shaping political attitudes and behavior.
THE PERSISTENCE OF BREXIT IDENTITIES
Given the intensity of the referendum campaign and the significance of the outcome, it is not surprising that British society became divided along Brexit lines at the time of the vote. More surprising has been the persistence of the social identities that were produced by the referendum (Hobolt, Leeper, and Tilley Reference Hobolt, Leeper and Tilley2021; Hobolt and Tilley Reference Hobolt and Tilley2021). Using a series of 18 surveys from the beginning of 2017 until October 2022, figure 1 shows the proportion of respondents who identified as a Leaver or a Remainer. There has been a drift downward since early 2020 in the number who identify with one side or the other of the Brexit divide. This has been more pronounced for the Leave side; with approximately 10% more Remainers than Leavers in 2022.Footnote 1 Nonetheless, even in October 2022—nearly three years after Britain had left the EU and more than six years after the referendum outcome—almost two thirds of people retained a Brexit identity.
Figure 1 Brexit Identities Over Time
The figure shows the percentage of respondents who “think of themselves as a Remainer or Leaver” using an 18-wave tracker survey run by YouGov.
These identities are both relatively stable over time and strongly held. We measured identity strength using a battery of five questions that gauge the emotional attachment to an identity. These questions were based on items used to test partisan attachment in the United States (Green, Palmquist, and Schickler Reference Green, Palmquist and Schickler2004; Huddy, Mason, and Aarøe Reference Huddy, Mason and Aarøe2015).Footnote 2 As shown in table 1, there was little variation over time in the mean score for either Brexit identity. It is notable that these attachments to Brexit identities are also stronger than attachments to British party identities. In fact, Brexit identity attachment is only slightly weaker than partisan attachment in the United States (Huddy, Mason, and Aarøe Reference Huddy, Mason and Aarøe2015, 7). A simpler measure of strength asked respondents how strongly they held their identity. In February 2022, 59% of Leavers and 61% of Remainers still stated that their identity was very or extremely important to them, compared to only 36% of Conservatives and 52% of Labour identifiers.
Table 1 Emotional Attachment to Brexit and Partisan Identities (1–5 Scale)
Notes: Emotional attachment uses five questions to measure respondents’ emotional attachment to their identity. Higher scores indicate greater attachment. Leavers, Remainers, Conservatives, and Labour were categorized by self-identity. Data for 2017 are from Hobolt, Leeper, and Tilley (Reference Hobolt, Leeper and Tilley2021); data from 2020–2022 are from the tracker survey run by YouGov.
These strongly held Brexit identities have shaped voters’ views. They generate affective polarization (Hobolt, Leeper, and Tilley Reference Hobolt, Leeper and Tilley2021) and perceptual biases (Pickup et al. Reference Pickup, de Rooij, van der Linden and Goodwin2021; Sorace and Hobolt Reference Sorace and Hobolt2021). These perceptual biases are most obvious in the economic realm. For example, at the end of 2021, 80% of Labour Remainers and 65% of Conservative Remainers believed the economy had worsened during the year; however, those numbers are only 53% for Labour Leavers and 45% for Conservative Leavers. Voters viewed the world through both a partisan and a Brexit lens. Of course, Brexit identities are based to some extent on existing divisions within the electorate (Clarke, Goodwin, and Whiteley Reference Clarke, Goodwin and Whiteley2017; Jennings and Stoker Reference Jennings and Stoker2017; Sobolewska and Ford Reference Sobolewska and Ford2020). However, these groups had neither clear labels nor subjective in-group identifications before the referendum. The referendum therefore defined and labeled in-groups based on the vote choice. The question is whether these new social identities also reflect a consistent set of policy preferences beyond views about the EU.
BREXIT IDENTITIES AND POLICY NORMS
Using a representative sample of the British electorate from July 2019, figure 2 shows differences in levels of agreement between Leavers and Remainers on various policy positions. These differences are shown separately for Labour and Conservative partisans to illustrate that while these Brexit identities overlap with partisanship, they also are separate from party identities. The policy positions are based on the well-established, two-dimensional set of ideological values developed in Britain by Heath, Evans, and colleagues in the 1990s (Evans, Heath, and Lalljee Reference Evans, Heath and Lalljee1996; Heath, Evans, and Martin Reference Heath, Evans and Martin1994) and cover the two main dimensions of political contestation in the United Kingdom and the rest of Europe: the economic left–right dimension and the social liberal–conservative dimension (Kriesi et al. Reference Kriesi, Grande, Lachat, Dolezal, Bornschier and Frey2008). The economic dimension includes attitudes toward redistribution, employment guarantees, nationalization (e.g., of utilities and public transport), and support for unions. The second “cultural” dimension includes attitudes toward criminal sentencing, the death penalty, censorship, immigration, and foreign aid. All of the answers were scored on a 1–5 scale from “strongly support” to “strongly against” (see appendix 1 for the full wording of the questions). The positive differences in figure 2 indicate that Remainers are more economically left-wing or more socially liberal than Leavers.
Figure 2 Differences in Policy Preferences Between Remainers and Leavers by Partisanship
The percentages show mean differences between Remainers and Leavers for policy attitudes on 1–5 scales, separately for Conservative and Labour partisans. Positive differences indicate that Remainers are more economically left-wing (on redistribution, employment guarantees, nationalization, and support for unions) and more socially liberal (on immigration, criminal sentencing, the death penalty, censorship, and foreign aid).
There were only very small differences between Leavers’ and Remainers’ economic left–right policy positions when they belonged to the same partisan in-group. Once we know someone’s party identity, there is no added value in knowing their Brexit identity when predicting their left–right values.Footnote 3 However, this was not the case for policies associated with the cultural dimension for which we see substantial differences between Remainers and Leavers, even within the same party group. Remainers are consistently more socially liberal than Leavers. These large differences apply to both Labour and Conservative partisans. The average difference between Leavers and Remainers who are Conservative partisans across all five of the social liberal–conservative questions was more than half a point on the 1–5 scales. For Labour partisans, that difference was more than 1 point. On immigration, Conservative Remainers were almost 1 point higher than Conservative Remainers, and Labour Remainers were almost 1.5 points higher than Labour Leavers.
Thus, it appears that the Brexit divide maps on to the cultural dimension of politics that has become increasingly salient in other Western democracies (Hooghe and Marks Reference Hooghe and Marks2018; Kriesi et al. Reference Kriesi, Grande, Lachat, Dolezal, Bornschier and Frey2008). This corroborates research that shows that this dimension is correlated with the Brexit vote choice (Hobolt Reference Hobolt2016; Surridge Reference Surridge2018). It also is not surprising given that education, age, and national-identity strength were all strong predictors of the EU referendum vote (Clarke, Goodwin, and Whiteley Reference Clarke, Goodwin and Whiteley2017; Curtice Reference Curtice2017; Evans and Tilley Reference Evans and Tilley2017; Hobolt Reference Hobolt2016; Sobolewska and Ford Reference Sobolewska and Ford2020) and, in Britain, correlate with people’s social liberalism (Surridge Reference Surridge2016; Tilley Reference Tilley2005). The extent to which these new identity groups have clear group norms about their in-group and out-group is less well understood, however.
To explore the existence of social norms for Brexit in-groups, figure 3 replicates figure 2 but uses people’s perceptions of the different groups. We asked separate representative samples about four different groups: (1) people who voted Remain in 2016 and normally vote Conservative; (2) people who voted Leave and normally vote Conservative; (3) people who voted Remain and normally vote Labour; and (4) people who voted Leave and normally vote Labour (see appendix 1 for full details). Figure 3 shows separately the perceptions of differences between Leavers and Remainers for perceived Conservative and Labour partisans. It is interesting that people appeared to have a reasonable idea of the reality shown in figure 2. There were no perceived differences between Leavers and Remainers on economic policy but substantial perceived differences in terms of social liberalism. These differences were, on average, approximately a third of a point and, again, were highest for immigration policy. People are able to identify policy norms for the two Brexit sides with a fair degree of accuracy.
Figure 3 Perceptions of Differences in Policy Preferences Between Remainers and Leavers
The percentages show mean differences between perceptions of Remainers and Leavers for policy attitudes on 1–5 scales, separately for perceived Conservative and Labour partisans. Positive differences indicate that Remainers are perceived to be more economically left-wing (on redistribution, employment guarantees, nationalization, and support for unions) and more socially liberal (on immigration, criminal sentencing, the death penalty, censorship, and foreign aid).
Figures 2 and 3 show that some policy preferences correlate with Brexit identity (i.e., there were differences in what Leavers and Remainers think) and that perceptions of the policy norms of the two Brexit identity groups were different (i.e., there were differences in what people think Leavers and Remainers think), but do Remainers and Leavers shift their views to match those policy norms? Do people want to think like their in-group?
According to the social identity literature, we would expect that people are keen to align their preferences with the norm of their in-group (Groenendyk, Kimbrough, and Pickup Reference Groenendyk, Kimbrough and Pickup2022; Sherif and Sherif Reference Sherif and Sherif1953; Suhay Reference Suhay2015). To examine this, we designed a survey experiment that randomly assigned people to information about the policy preferences of Brexit identity groups to examine the effects on their own preferences (Tilley and Hobolt Reference Tilley and Hobolt2023b). Table 2 shows the results of this experiment, fielded in January 2020 (see appendix 2 for the full details). After asking respondents their Brexit identity, we gave those in the treatment group truthful information on their in-group’s and out-group’s policy preferences for immigration policy. Specifically, we told respondents in the treatment group the following:
There is a lot of debate about reducing immigration to Britain at the moment. A recent survey has shown that most Leavers agree that immigration should be reduced, whereas most Remainers do not agree that immigration should be reduced. To what extent do you agree that immigration should be reduced?
Table 2 Mean Support for Pro-Immigration Policy (1–5 Scale)
Notes: *p<0.05. The dependent variable is a 1–5 scale measuring respondents’ disagreement with the statement that “immigration should be reduced.” High scores indicate greater support for a pro-immigration policy. Leavers and Remainers were categorized by self-identity.
There were five response options to the question from “strongly agree” to “strongly disagree.” Those in the control group were asked, with the same response options, simply: “There is a lot of debate about reducing immigration to Britain at the moment. To what extent do you agree that immigration should be reduced?” Table 2 shows the mean scores for Remainers and Leavers by treatment. Higher scores indicate greater support for immigration. Considering the control group first, it is worth noting that we found the same large differences between Remainers and Leavers in their views on immigration policy. As in figure 2, Remainers were more in favor of immigration than Leavers, and the difference again was substantial: 1.5 points on a 1–5 scale. More interestingly, we also observed effects of the group norms treatment. Specifically, Remainers in the treatment group were more pro-immigration than Remainers in the control group by a quarter of a point (statistically significant at the 5% level). Remainers thus appeared to shift their policy positions to conform with their in-group and reject their out-group.
We did not observe a similar effect for Leavers, however. One possible explanation for this is the likely presence of floor effects for Leavers when it comes to this policy area. Of Leavers, 51% strongly agreed (scoring 1) and 35% agreed (scoring 2) with the policy of reducing immigration. In that sense, there was probably little room for group cues to affect people’s own views.
CONCLUSION
The Brexit referendum left Britain a politically divided country. Previous research has shown that Leavers and Remainers became new political identities that were stable over time and strongly held, even after Britain left the EU (Curtice Reference Curtice2018; Hobolt, Leeper, and Tilley Reference Hobolt, Leeper and Tilley2021). Six years later, these identities have weakened somewhat, primarily due to a reduction in the number of Leavers, but both identities continue to survive in post-Brexit Britain. Furthermore, our study shows that these new identities are associated with policy norms that go beyond views on European integration. Leavers and Remainers are clearly on opposite sides of the cultural dimension. In fact, people are not only aware of these differences among Leavers and Remainers; they also appear to align themselves with the preferences of their in-group. This helps to explain why Brexit identities have been so stable and strongly held even in the absence of clear partisan cues. This also suggests that although the Brexit issue lost its salience during the COVID-19 pandemic, the political schism triggered by Brexit could be remobilized because it is rooted in deeper value divides and norms.
Leavers and Remainers are clearly on opposite sides of the cultural dimension. In fact, people are not only aware of these differences among Leavers and Remainers; they also appear to align themselves with the preferences of their in-group.
Our findings also raise a more general point about social identities. The aftermath of the EU referendum in Britain demonstrates that politically salient and distinct group identities can emerge over a relatively short time and then persist. These new identities can rival partisan identities in terms of stability, strength of feeling, and even shared policy norms. Although some people have changed their mind about Brexit, the vast majority of those who voted in the referendum and identified with one side of the debate have remained true to their chosen in-group. Moreover, these identities have continued to shape people’s outlook on politics long after the referendum. This suggests that political identities with shared policy norms are not unique to partisanship but that highly visible events (e.g., the Brexit referendum) can also create new identities, thus activating value divides that crosscut existing partisan divisions.
ACKNOWLEDGMENTS
This research was generously supported by the Economic and Social Research Council (Grant Nos. ES/R000573/1 and ES/V004360/1). We are grateful to the symposium editor, David Leal, and the anonymous reviewers for their helpful comments. We also thank Mark Pickup for his insightful suggestions on earlier versions of this article.
DATA AVAILABILITY STATEMENT
Research documentation and data that support the findings of this study are openly available at the PS: Political Science & Politics Harvard Dataverse at https://doi.org/10.7910/DVN/X08IVC.
CONFLICTS OF INTEREST
The authors declare that there are no ethical issues or conflicts of interest in this research.
APPENDIX 1
The below questions are used to measure the two main dimensions of policy preferences: left-right economic preferences (1–4 below) and socially liberal versus socially conservative preferences (5–9 below). They were fielded to a representative sample of the British electorate by YouGov in July 2019 (N=1,758). The question order was randomized before being presented to respondents. All policy areas had the same response options of strongly support, support, neither support nor against, against, strongly against.
‘We’d like to ask you how much you support the following:
-
1. Strong trade unions to protect employees’ working conditions and wages
-
2. Making sure that major public services and industries are in state ownership
-
3. The government providing a job for everyone who wants one
-
4. Redistributing income from the better-off to those who are less well-off
-
5. Spending less on foreign aid
-
6. Censorship of films and magazines to uphold moral standards
-
7. The death penalty
-
8. Giving stiffer sentences to people who break the law
-
9. Fewer immigrants being let into the country’
Respondents were also asked their partisan and Brexit identity prior to the policy positions using the following two questions:
‘Generally speaking, do you think of yourself as Labour, Conservative, Liberal Democrat or what? Conservative; Labour; Liberal Democrat; Scottish National Party (SNP); Plaid Cymru; United Kingdom Independence Party (UKIP); Green Party; Other party; None’
‘Since the EU referendum, some people now think of themselves as Leavers and Remainers, do you think of yourself as: a Leaver; a Remainer; neither a Leaver nor a Remainer; don’t know’
The below questions were used to measure perceptions of policy norms, again with the same policy areas as listed above, and the same response options of strongly support, support, neither support nor against, against, strongly against. Each battery of norms questions was asked to a separate representative sample of the electorate (between 811 and 878 people in each of the four groups) by YouGov in July 2019.
‘We are now going to ask you about what people who voted Leave in the 2016 EU referendum support and normally support the Conservatives think about a range of issues. How much do you think people like that would support the following:’
‘We are now going to ask you about what people who voted Remain in the 2016 EU referendum support and normally support the Conservatives think about a range of issues. How much do you think people like that would support the following:’
‘We are now going to ask you about what people who voted Leave in the 2016 EU referendum support and normally support Labour think about a range of issues. How much do you think people like that would support the following:’
‘We are now going to ask you about what people who voted Remain in the 2016 EU referendum support and normally support the Labour think about a range of issues. How much do you think people like that would support the following:’
APPENDIX 2
The policy norm experiment was run as a survey-embedded vignette experiment in January 2020 by YouGov, using a representative sample of the British electorate. Respondents were randomly allocated to a control or treatment group. The control group (N=451) were asked:
‘There is a lot of debate about reducing immigration to Britain at the moment. To what extent do you agree that immigration should be reduced?’
The response options were strongly agree, agree, neither agree nor disagree, disagree, strongly disagree. The treatment group (N=431) were given extra (correct) information on policy norms before the same question. The treatment group was thus asked:
‘There is a lot of debate about reducing immigration to Britain at the moment. A recent survey has shown that most Leavers agree that immigration should be reduced, whereas most Remainers do not agree that immigration should be reduced. To what extent do you agree that immigration should be reduced?’
Both the control and treatment groups were asked prior to the immigration question their Brexit identity using the below question:
‘Since the EU referendum, some people now think of themselves as Leavers and Remainers, do you think of yourself as: a Leaver; a Remainer; neither a Leaver nor a Remainer; don’t know’
The control group contained 219 Leavers, 199 Remainers, and 153 people who answered neither or don’t know. The treatment group contained 200 Leavers, 221 Remainers, and 143 people who answered neither or don’t know.
Open access
