ArticlePDF Available

Divided by the Vote: Affective Polarization in the Wake of the Brexit Referendum


Abstract and Figures

A well-functioning democracy requires a degree of mutual respect and a willingness to talk across political divides. Yet numerous studies have shown that many electorates are polarized along partisan lines, with animosity towards the partisan out-group. In this article, we further develop the idea of affective polarization, not by partisanship, but instead by identification with opinion-based groups. Examining social identities formed during Britain’s 2016 referendum on European Union membership, we use surveys and experiments to measure the intensity of partisan and Brexit-related affective polarization. The results show that Brexit identities are prevalent, felt to be personally important, and cut across traditional party lines. These identities generate affective polarization as intense as that of partisanship in terms of stereotyping, prejudice, and various evaluative biases, convincingly demonstrating that affective polarization can emerge from identities beyond partisanship.
Content may be subject to copyright.
Divided by the vote: aective polarization in the wake of the Brexit
LSE Research Online URL for this paper:
Version: Accepted Version
Hobolt, Sara, Leeper, Thomas J. and Tilley, James (2020) Divided by the vote:
a*ective polarization in the wake of the Brexit referendum. British Journal of
Political Science. ISSN 0007-1234 (In Press)
Items deposited in LSE Research Online are protected by copyright, with all rights
reserved unless indicated otherwise. They may be downloaded and/or printed for private
study, or other acts as permitted by national copyright laws. The publisher or other rights
holders may allow further reproduction and re-use of the full text version. This is
indicated by the licence information on the LSE Research Online record for the item.
Divided by the Vote:
Affective Polarization in the Wake of the Brexit Referendum
Sara B. Hobolt, Thomas J. Leeper and James Tilley
**forthcoming in the British Journal of Political Science**
A well-functioning democracy requires a degree of mutual respect and a willingness to talk
across political divides. Yet numerous studies have shown that many electorates are polarized
along partisan lines, with animosity towards the partisan out-group. In this article, we further
develop the idea of affective polarization, not by partisanship, but instead by identification with
opinion-based groups. Examining social identities formed during Britain’s 2016 referendum
on European Union membership, we use surveys and experiments to measure the intensity of
partisan and Brexit-related affective polarization. The results show that Brexit identities are
prevalent, felt to be personally important, and cut across traditional party lines. These identities
generate affective polarization as intense as that of partisanship in terms of stereotyping,
prejudice, and various evaluative biases, convincingly demonstrating that affective polarization
can emerge from identities beyond partisanship.
Key words: Affective polarization; Partisanship; Group identity; Brexit; Bias; Referendum
In recent years, scholars of American democracy have pointed to growing affective
polarization along partisan lines. Republicans and Democrats have developed strong
emotional attachments toward co-partisans and hostility toward opposing partisans (Iyengar
et al. 2012; Iyengar and Westwood 2015; Mason 2015, 2018). This is worrying as a well-
functioning democracy requires that citizens and politicians are willing to engage respectfully
with each other, even on controversial topics (Lipset 1959; Dahl 1967). Where we see instead
mass affective polarization, we find intolerance and political cynicism (Layman et al. 2006)
and reduced opportunities for collaboration and compromise (MacKuen et al. 2010). But is
affective polarization limited to partisanship? In this article we argue that such polarization
can emerge along lines drawn not just by partisan loyalties, but also by identification with
opinion-based groups. We thus aim to significantly expand the scope of identities and
political contexts that might be examined through the lens of affective polarization. Building
on theories of social identity, we argue that significant political events can generate affective
polarization. They do this by causing people to identify with others based on a shared opinion
about the event. We study this phenomenon of opinion-based group identities in the wake of
a critical juncture in British politics: the 2016 referendum on Britain’s European Union (EU)
membership. Our data suggest that affective polarization is a phenomenon not unique to
partisanship, and that animosity across opinion-based groups can cross-cut longstanding
partisan divisions.
We make three significant contributions. First, we present an original conceptualization of
affective polarization based on an opinion-based in-group identity that focuses on three
different core components: identification with an in-group based on a common cause,
differentiation from the out-group leading to prejudice and animosity, and evaluative bias in
perceptions of the world and in decision-making. Second, we examine this phenomenon
empirically, using evidence from a large and diverse range of existing data, original surveys,
and novel experiments. We demonstrate the scope of affective polarization after the Brexit vote
using implicit, explicit, and behavioural indicators. Finally, we directly compare the impact of
these new opinion-based Brexit identities to traditional partisan divisions. We find a similar
degree of affective polarization for the new Brexit identities as for party identities in terms of
identification, differentiation, and evaluative bias. Moreover, Brexit identities cut across
traditional party lines meaning that affective polarization is neither restricted to partisanship,
nor a mere proxy for partisan affect. We argue that these new identities reflect pre-existing, but
less-politicized, social divisions, like age and education, which were mobilized in the context
of the referendum and have consolidated into the newly salient identities: ‘Leave’ and
‘Remain’. These findings have important implications for the study of social identities and
electoral democracy, not least because they demonstrate the emergence of strongly held
political identities over a relatively short period of time.
The article proceeds as follows. We discuss the literature on in-group identities and affective
polarization and present our conceptualization of opinion-based group identities. We then
briefly introduce the context of the referendum, and proceed to show evidence of identification
with the in-group, differentiation towards the out-group, and evaluative biases for both Brexit
and partisan identities. All three effects are at least as large, if not larger, for Brexit identity
compared to partisan identity. In conclusion, we discuss the sustainability of opinion-based
cleavages and consider the conditions in which polarization along these lines is triggered.
Affective polarization and opinion-based groups
Inherent in all democratic systems is the constant threat that the group conflicts which are
democracy’s lifeblood may solidify to the point where they threaten to disintegrate society.
Seymour Martin Lipset (1959, 83).
Political conflict and competition are at the heart of democratic life (Schattschneider 1960).
The classic ideal of democracy is not one absent of conflict, but rather one where a single
conflict is not so entrenched and all-encompassing that society suffers (Dahl 1967). As the
quotation from Lipset highlights, the health of democracy is threatened when conflicts solidify
and political identities crystallize into polarized groups. At its most extreme, we see ethnically
divided societies where government-opposition dynamics are almost entirely replaced by
‘ethnic outbidding’ (Rabushka and Shepsle 1972) and where the democratic opposition is seen
as ‘the enemy of the people’ by those in power (Horowitz 1993).
But mass polarization can also occur in societies not plagued by such divisions. The most
prominent example is the increasing partisan polarization in American politics over the last
few decades. While there remains some debate about the particular form of polarization at the
mass level (Fiorina and Abrams 2008), there is a broad consensus that the US public has
become more divided along partisan and ideological lines in recent years (Hetherington 2009;
Layman et al. 2006; Mason 2018). Most notably, there has been rising interpersonal animosity
across party lines, with Democrats and Republicans increasingly expressing dislike for one
another (Layman et al. 2006; Iyengar et al. 2012; Iyengar and Westwood 2015; Mason 2015,
2018). This phenomenon has been described as affective polarization, defined as an emotional
attachment to in-group partisans and hostility towards out-group partisans (Green et al. 2004;
Iyengar et al. 2012, 2019; Iyengar and Westwood 2015). While affective polarization is often
rooted in policy disagreement, it is distinct from ideological polarization. The latter concerns
the extremity of political views, whereas the former is focused on hostility towards outgroups
(Iyengar et al. 2012; Mason 2015, 2018). In other words, affective polarization does not
necessarily imply extreme policy disagreement. Studies on affective polarization in the US
have shown that antipathy towards partisan opponents has escalated substantially among
voters. This has meant that increased in-party favouritism has been matched by greater negative
stereotyping and out-group discrimination (Iyengar and Westwood 2015; Lelkes and
Westwood 2017; Mason 2013, 2015, 2018; Miller and Conover 2015).
There are many worrying consequences of affective polarization. Out-group animosity makes
it more difficult for citizens to deliberate without prejudice and to seek diverse perspectives on
controversial topics (Valentino et al. 2008). This in turn impairs democratic dialogue,
collaboration, and compromise (MacKuen et al. 2010) and may lead to the erosion of trust in
political institutions and the democratic legitimacy of elected leaders (Layman et al. 2006;
Anderson et al. 2005). Affective polarization also exacerbates ‘filter bubbles’ and ‘echo
chambers’ as people become unwilling to engage (in person or online) with people from the
other side (Levendusky 2013; Levendusky and Malhotra 2016.)
The concept of affective polarization is rooted in social psychological research on social
identity and intergroup conflict, most prominently work on social identity theory by Henri
Tajfel (Tajfel 1970, 1979; 1982; Tajfel and Turner 1979). The core idea is that group
membership is an important source of pride and self-esteem. It gives each of us a sense of social
identity. Yet it also means that our sense of self-worth is heightened by discriminating against,
and holding prejudiced views of, the out-group (Tajfel 1970; Tajfel 1979). According to Tajfel
and Turner (1979), there are three mental processes involved in shaping a social identity: social
categorization, in which we distinguish between ‘us’ and ‘them;’ social identification, in which
we adopt the identity of the group we have categorized ourselves as belonging to; and social
comparison, in which we compare our own group favourably to others. This desire to compare
oneself with an out-group often, although not always, creates competitive and antagonistic
intergroup relations. This then serves to further heighten identification with the in-group.
While social identity theory has proved extremely useful to political science (for an excellent
review see Huddy (2001)), the identities considered, such as race, gender, and partisanship,
have been the same social categories common to psychological research (Tajfel and Turner
1979; Mason 2015). Partisanship has been particularly central, after all ‘in the political sphere,
the most salient groups are parties and the self-justifications that sustain group life are primarily
grounded in - and constructed to maintain - partisan loyalties’ (Achen and Bartels 2016, 296).
Less attention has been paid to other political identities,1 even though self-categorized social
identities are inherently subjective (Turner 1982; McGarty et al. 2009). We argue that affective
polarization can also stem from political identities defined by shared political opinions. Our
argument builds on a recent strand in the social psychology literature that has developed the
notion of opinion-based groups (Bliuc et al. 2007; McGarty et al. 2009). Merely holding the
same opinion as others is not sufficient for such a group to exist, rather the shared opinion
needs to become the basis of a social identity. In other words, people need to define themselves
in terms of their opinion group membership in the same way as they would any other
meaningful social group, such as a religious denomination or political party. Opinion-based
groups emerge in the context of salient inter-group comparisons: that is, situations where
people are compelled to take sides on an issue. Research suggests such identities may emerge,
or crystallize, in response to dramatic events, such as wars or man-made disasters (McGarty et
1 There are two notable exceptions to this. The first is self-identification as a conservative or a liberal in the US.
This has been shown to function as a social identity which is separable from issue positions (Malka and Lelkes
2010; Mason 2018; Kinder and Kalmoe 2017). The second is gun ownership in the US which Lacombe (2019)
shows to be a distinct social identity which shapes political action.
al. 2009; Smith et al. 2015). We argue they can also emerge from politically engineered events,
specifically referendums on political issues.
We conceptualize affective polarization of opinion-based groups as having three necessary
components: (1) in-group identification based on a shared opinion; (2) differentiation of the in-
group from the out-group that leads to in-group favourability and out-group denigration; and
finally, (3) evaluative bias in perceptions of the world and in decision-making. The starting
point of affective polarization is that individuals must have internalized their group
membership as an aspect of their self-identification. People form a social identity (Tajfel and
Turner 1979), but in this case it is based on group membership due to a common cause
(McGarty et al. 2009), rather than organized around a social category. Similar to partisanship,
‘people think of themselves as members of a group, attach emotional significance to their
membership and adjust their behavior to conform to group norms’ (Bartle and Belluci 2009, 5;
see also Klar 2014; Westwood et al. 2018).2 The next step is that people must favourably
compare their own group with the out-group (Tajfel and Turner 1979). This differentiation
therefore means that a second indicator of affective polarization is prejudice towards and
stereotyping of members of the out-group. The final step is that group competition must also
spill over into perceptions and political and non-political decision-making. When it comes to
opinion-based polarization, in-group bias will be an omnipresent feature that affects opinions
and decision-making in ways that go beyond the specific in-group conflict. People will evaluate
political outcomes via the lens of their identity and make decisions based on that identity. To
diagnose affective polarization, we should therefore observe all three of these factors:
2 Equally, Mason (2015) demonstrates partisan social polarization in terms of our second and third components:
affect (anger) towards the out-group, judgement (bias) of the out-group and behaviour (activism) towards the out-
identification, differentiation, and evaluative bias. In the remainder of the paper, we examine
these different aspects of affective polarization across opinion-based group membership in the
context of the 2016 referendum on Britain’s EU membership.
The 2016 Brexit Referendum
On the 23rd June 2016, British voters were asked in a nationwide referendum: ‘Should the
United Kingdom remain a member of the European Union or leave the European Union?’.
Despite the ‘Remain’ side having the endorsement of all the major parties in Parliament,3 52
per cent of the British electorate voted to exit the EU (‘Brexit’). This sent shockwaves through
Britain and Europe. Never before had a member state decided to leave the European Union.
Although the vast majority of British parliamentarians voted to trigger the Brexit negotiation
process, and both major parties campaigned on a platform of taking the UK out of the EU in
the 2017 general election, the public did not universally rally behind Brexit. As we will show,
the referendum and campaign triggered affective polarization over the issue of leaving or
remaining in the EU that continued to divide society. Perhaps surprisingly, this occurred even
though the question of EU membership and European integration was not a highly salient issue
to the electorate before the referendum. During the 2015 General Election, only a year ahead
of the referendum vote, less than 10 per cent of people identified the EU as the among the two
most important issues facing Britain,4 and the issue of the EU played a minimal role in the
election campaign. Prior to the referendum, Britain’s role in the European Union was not a
3 The governing Conservative Party was openly divided with several cabinet members campaigning to leave the
EU, however. Some high-profile members of the Labour Party also endorsed Leave (Hobolt 2016; Evans and
Menon 2017).
4 See IPSOS Mori (2018) for time series data on the question: ‘What would you say is the most important issue
facing Britain today? What do you see as other important issues facing Britain today?’.
highly salient political issue, let alone a social identity, among voters. The opinion-based group
identities ‘Leaver’ and ‘Remainer,’ which we will show came to take on considerable meaning
for most British voters, have no long-term history in British politics. There were no labels for
sides in the Brexit debate until the campaign itself.5
The aftermath of the Brexit referendum is thus an apt case for the study of affective polarization
around opinion-based groups. Social identity theory suggests that salient group identities
emerge when people are compelled to take sides in a debate. A referendum that asks people to
take a stance in favour (Leave) or against (Remain) exiting the EU, is such a case. Moreover,
the question of leaving the EU is unusual in that it cut across traditional party lines meaning
that the divisions resulting from the referendum were not immediately subsumed into the
existing party divide. Yet, while a large body of literature has examined the determinants of
voting behaviour in the referendum (Goodwin and Heath 2016; Hobolt 2016; Becker et al.
2017; Clarke et al. 2017; Colatone and Stanig 2018), we know much less about the way in
which the vote subsequently divided people.
5 Indeed, ‘Leave’ and ‘Remain’ were seemingly innocuous labels created by a decision of the UK Electoral
Commission in September 2015 to improve the intelligibility of the referendum question (Electoral Commission
2015) which had originally been worded ‘Should the United Kingdom remain a member of the European Union?’
with the options ‘Yes’ and ‘No.’ One possible precursor identity for the Leave side is ‘Euro-sceptic’. However
data from the 2005 British Social Attitudes survey shows that even when prompted only 15 per cent of people
thought of themselves in this way. This seems relatively low given the question context in which 23 per cent of
people thought of themselves as environmentalists, 8 per cent as anti-war campaigners and 9 per cent as animal
rights campaigners. Euro-sceptic was not a strongly held identity in 2005 either. Only 20 per cent of that 15 per
cent (i.e. 3 per cent of the population) thought that they had ‘a lot more in common’ with fellow Euro-sceptics.
To empirically examine affective polarization in the context of Brexit we use multiple sources
of survey and experimental data. All of our data comes from public opinion surveys that are
designed, and further weighted, to be representative of the British population. Table 1 presents
an overview of these datasets. We rely upon both the largest existing data source on public
attitudes toward the referendum, namely the British Election Study 2016-2019 panel
(Fieldhouse et al. 2019), as well as a series of original public opinion surveys and survey
experiments conducted between 2017 and 2019. Most of these surveys were conducted by
YouGov, a prominent polling organization that uses quota sampling and reweighting methods
to generate nationally representative samples from an online, opt-in pool of over 1 million
British adults. We also supplement these data further with surveys from Sky Polling, which
applies similar methods to a panel consisting of subscribers to the widely used Sky satellite
television service.6 This variety of data sources means that all our results come from nationally
representative samples, but are not dependent on any single data source or survey methodology.
Given the number and diversity of research designs and measures deployed, we describe each
alongside its results in what follows.
6 Approximately 12 million UK households (44 per cent) have a Sky subscription.
Table 1. Data sources
Full description
Time period
Question coverage
British Election
Study multi-wave
panel survey
9 waves from
April 2016 to
March 2019
Party identity, Brexit identitya
and emotional attachment to
both identities
~30,000 per
YouGov repeated
9 waves from
April 2017 to
September 2019
Brexit identity b, economic
perceptions (Jan 2018) and
party identity (Jan 2018)
(~1,700 per
YouGov cross-
sectional survey
September 2017
Party identity, Brexit identityb,
prejudice and perceptions for
both identities
Sky Polling cross-
sectional surveys
October and
November 2017
Party identity, Brexit identityb,
emotional attachment and
prejudice for both identities
YouGov conjoint
experiment about
choice of BBC
October 2017
Out-group prejudice and in-
group bias
YouGov conjoint
experiment about
choice of a lodger
October 2017
Out-group prejudice and in-
group bias
a. Question asks whether respondent thinks of themselves as ‘closer to the either the Leave or Remain side’
b. Question asks whether respondent thinks of themselves as a Leaver or Remainer.
Note: All survey respondents are drawn from online panels involving quota sampling, which are then weighted to
be representative of the British population with respect to demographic characteristics.
As we argued above, there are three key components of affective polarization along opinion-
based lines: in-group identification, group differentiation (especially prejudice towards
members of the out-group), and evaluative bias in both perceptions and decision-making. We
thus begin by examining the prevalence of Brexit identities in the electorate using the BES,
YouGov, Sky, and Tracker surveys as well as the strength and importance of these identities
using the BES and Sky surveys. Next, we examine how those with Leaver and Remainer
identities stereotype those on each side of the divide and the extent to which they display
prejudice against their Brexit out-group using Sky and YouGov surveys. Then we show how
these identities colour citizens’ perceptions of economic performance in a manner that cross-
cuts partisan identities. Finally, we measure the degree to which Brexit identities shape
judgements of political and non-political choices using revealed choice conjoint experiments.
Our starting point is simply to measure the proportion of people willing to express an identity
linked to the referendum. Table 2 shows two ways of measuring Brexit identity. The question
included in the YouGov and Sky surveys asks people: ‘Since the EU referendum last year,
some people now think of themselves as Leavers and Remainers, do you think of yourself as a
Leaver, a Remainer, or neither a Leaver or Remainer?’. This mirrors the standard party identity
question which asks people: ‘Generally speaking, do you think of yourself as Labour,
Conservative, Liberal Democrat or what?’7 The BES survey uses a slightly different format
which does not mention the two identity labels and encourages people to pick a side: ‘In the
EU referendum debate, do you think of yourself as closer to either the Remain or Leave side?’
7 A list of parties is then provided to respondents which, as well as Conservative, Labour and Liberal Democrat,
also includes the SNP (in Scotland only), Plaid Cyrmu (in Wales only), UKIP, the Greens, the BNP, ‘other’ party
and none. For both party and Brexit identity, people are given a ‘don’t know’ option. We have coded this as
equivalent to no identity.
Table 2. Comparison of the strength of party and Brexit identities
Party identity
Brexit identity
Proportion of people with identity
BES - June 2017
YouGov - Sep 2017
Sky – Oct/Nov 2017
Brexit identity scale (1-5 scale of 5-question battery)
BES - June 2017
Sky – Oct/Nov 2017
a. Question asks whether respondent thinks of themselves as ‘closer to the either the Leave or Remain side’, rather
than whether they think of themselves as a Leaver or Remainer.
Note: The BES data have a total unweighted N of 31,197. The YouGov data have a total unweighted N of 3,326.
The Sky data have a total unweighted N of 1,692 for party identity and 1,702 for Brexit identity. The emotional
attachment scale consists of five questions (with a 1-5 Likert scale) that ask respondents with an identity whether
a) they talk about ‘we’ rather than ‘they’, b) criticism of their side feels like a personal insult, c) they have a lot in
common with people on their own side, d) they feel connected with other supporters of their own side and e) they
feel good when people praise their own side. High scores indicate greater agreement. These questions were only
asked of those with a relevant political identity.
As Table 2 shows the BES data gave high proportions of people with a Brexit identity (over 85
per cent). Yet, even with the weaker wording on the YouGov and Sky surveys about three
quarters of people identified themselves as Leavers or Remainers. This is despite these surveys
being conducted over 18 months after the actual referendum. Unsurprisingly given the close
referendum vote, there are roughly even numbers of Leavers and Remainers.8 The total number
of people with a Brexit identity looks similar to the proportions of people who identify with a
party. For example, the YouGov data show that 57 per cent of people identified with one of the
two main parties. Not shown are another 16 per cent of people who identified with one of the
other minor parties. In total, 74 per cent of people had a party identity compared to 75 per cent
8 This is also the case when slightly different labels are used. Richards and Heath (2017) asked people in July
2017 whether they considered themselves a ‘Remainer’ or a ‘Brexiteer’. They find that 45 per cent of people
identified as Remainers and 42 per cent identified as Brexiteers.
with a Brexit identity in the YouGov data.9 The prevalence of Brexit identities and traditional
party identities is very similar. If anything, Brexit identities have become more widespread
than partisan identities.
The bottom half of Table 2 also shows that both types of identities are equally strongly held.
We show a measure of emotional attachment to people’s own identity using a battery of five
questions. These questions create a similar scale to that used by others (see Huddy et al. 2015;
Greene 2000; Green et al. 2004) and ask people whether they agree or disagree with the
following with regard to their own identity:
When I speak about the [respondent identity] side, I usually say “we” instead of
When people criticize the [respondent identity] side, it feels like a personal insult
I have a lot in common with other supporters of the [respondent identity] side
When I meet someone who supports the [respondent identity] side, I feel
connected with this person
When people praise the [respondent identity] side, it makes me feel good
Response options for all items were ‘strongly disagree’ to ‘strongly agree,’ scored 1-5 and
averaged.10 For the two main party identities the average score is around the midpoint of the
9 As we might expect, these identities overlap to some extent. Nonetheless, the Tracker survey in January 2018
showed that only 35 per cent of Leavers were Conservative identifiers, and only 36 per cent of Remainers were
Labour identifiers. Table A1 in the appendix shows this breakdown in more detail.
10 The BES data uses a 4-point scale, but don’t know responses are coded as 3 to make it analogous to the Sky
data. The items form very reliable scales for both identities. The Cronbach’s Alpha scores for the BES data are
.84 for Leavers, .85 for Remainers, .81 for Conservatives and .83 for Labour identifiers. Alpha scores for the Sky
data are .79 for Leavers, .74 for Remainers, .86 for Conservatives and .80 for Labour identifiers.
scale for both datasets. Interestingly this is not much lower than the scores for responses to
similar questions asked in the US (Green et al. 2004, 38; Huddy et al. 2015, 7). More
importantly for our purposes, these emotional attachment scores are slightly higher for Brexit
identities than they are for party identities. This is especially obvious for the Sky data which
use the Brexit identity question which is most analogous to the party identity question.
Overall, Table 2 reveals that not only were slightly more people willing to claim a Brexit
identity than a party identity, but the attachment that people had to that Brexit identity was, if
anything, slightly stronger than their party identity. Moreover, these Brexit identities appear to
be largely stable at the aggregate level. Figure 1 shows the numbers of people with a Brexit
identity over time for nine waves of the BES from April 2016 until March 2019 and for nine
waves of the Tracker survey from April 2017 until September 2019. Whether measured using
the BES closeness question or the Tracker identity question, the numbers of people with an
identity are almost completely static over time. Around three-quarters of people in Britain think
of themselves as Leavers or Remainers and this has been the case since the beginning of the
referendum campaign in early 2016 through to today. Most importantly there is aggregate level
stability in the numbers within each identity grouping, suggesting the same kind of unmoving
affective identity as partisanship. About half of those with an identity are Leavers and half are
Remainers, no matter what month we choose. These proportions have changed very little since
the referendum result. Indeed the small increase in the number of Remainers is almost entirely
due to an increased prevalence of that identity among people who did not, or were not able to,
vote in 2016.11
11 As the BES data are a repeated panel we can also look at the proportion of people who move in and out of an
identity over time. These numbers look very similar for party and Brexit identities. For example, 81 per cent of
people have the same party identity in June 2017 as they did in July 2016, whereas 87 per cent of people have the
Figure 1. Brexit identities over time
Note: The British Election Study asks whether the respondent thinks of themselves as ‘closer to the either the
Leave or Remain side’ and includes nine waves from April 2016 to March 2019. The Tracker data comprise nine
cross-sectional surveys from April 2017 to September 2019 and asks whether people think of themselves as a
Remainer or a Leaver.
same Brexit identity in June 2017 as they did in July 2016. In both cases most of the movement is from, and into,
no identity, rather than movement between different identities.
2016 2017 2018 2019 2020
% of people with identity
British Election Study
2016 2017 2018 2019 2020
% of people with identity
Tracker survey
Of course, while Brexit identities are a new part of British politics, they could reflect underlying
societal divides that predate the referendum. Research into the determinants of the Brexit vote
indicate that the referendum mobilized an underlying fault line between social liberals with
weak national identities, who tend to be younger and have more educational qualifications, and
social conservatives with stronger national identities, who tend to be older with fewer
educational qualifications (Hobolt 2016; Clarke et al. 2017; Evans and Tilley 2017; Jennings
and Stoker 2017; Curtice 2017a). Using BES data, Appendix 2 in the Supplementary
Information confirms that the key socio-economic predictors of a Leave identity relative to
Remain identity are age and education. By contrast, measures of social class (such as income,
occupation and housing tenure) continue to matter more for partisan identities than for Brexit
identities despite sharp falls in class voting in Britain in recent decades (Evans and Tilley
2017).12 Analysis of a subset of the BES data in Appendix 2 also confirms that people with
stronger British identities are more likely to hold a Leave identity, although the effect is not
huge. These correlates of Brexit identity are clearly important, but in this paper we are primarily
interested in how such political divides manifest themselves as social identities that facilitate
affective polarization. Whether the social and political forces driving diverging preferences
about European integration are new or not, the labels provided by the referendum campaign
certainly are. It is these labels that allow people to self-identify as a member of one opinion-
based group or the other. It is also these labels that allow for differentiation, favouritism
towards the in-group, and animosity towards the out-group.
12 The importance of education as a predictor of Brexit identity links to the rise of the cultural dimension in politics
across Europe. Divides along transnational integration–demarcation dimensions (Kriesi et al. 2006, 2008; Hooghe
and Marks 2018) are increasingly salient elsewhere in Europe. This dimension is distinct from the traditional
economic left–right dimension and is focused more on identity and cultural concerns.
For the emergence of Brexit identities to constitute affective polarization, we expect to see
Leavers and Remainers stereotype their in-group and out-group and express animosity toward
the out-group. Figure 2 shows people’s perceptions of their own and the other side in terms of
three positive personal characteristics (intelligent, open-minded and honest) and three negative
personal characteristics (selfish, hypocritical and closed-minded). This list of traits is similar
to that used by Iyengar et al. (2012) to examine partisan affective polarization over time and
space. Respondents were asked how well they thought these six different characteristics
described the two sides on a five-point scale from ‘not at all well’ to ‘very well’. ‘Very well’
is scored 5 and ‘not at all well’ is scored 1. We focus on both differentiation along partisan
lines, as a baseline, and differentiation along the lines of Brexit identity.
The top two graphs in Figure 2 show mean perceptions by party identity. We see a familiar
story. Perceptions of Conservative supporters, graphed on the left, are very different for people
who are themselves Conservative identifiers compared to those who are Labour identifiers.
Conservative partisans score their in-group above 3.5 in terms of intelligence, honesty and
open-mindedness, but are much more reluctant to say that their in-group are selfish,
hypocritical or closed-minded. The exact opposite is true for Labour partisans who score
Conservative supporters at nearly 4 in terms of their selfishness, hypocrisy and
closedmindedness, but are extremely unlikely to say that Conservatives might be intelligent,
open-minded or honest. The top right-hand graph shows perceptions of Labour supporters.
Again, Labour identifiers only attribute positive characteristics to their in-group while
Conservative identifiers only attribute negative characteristics to their out-group.
Fascinatingly, we see the very same patterns for Brexit identities in the two bottom graphs.
Remainers and Leavers are much more likely to attribute positive characteristics to their own
side and negative characteristics to the other side. The magnitude of these differences is very
large. Remainers’ average score for the three positive characteristics about their own side is
3.9 while their average score for the three negative characteristics about their own side is just
1.9. The gulf between agreement with negative and positive attributes of the out-group is also
huge. For Remainers’ perceptions of Leavers, the average score for the three positive
characteristics is 2.4, yet the average score for the three negative characteristics is 3.6. Nor are
these views of Leavers and Remainers driven by party identity. Appendix 3 contains four OLS
regressions that predict whether people have positive and negative views of both sides using
both party identity and Brexit identity. All four models show very large effects of Brexit
identity and very weak effects of party identity on perceptions of Remainers and Leavers.
Figure 2. Perceived characteristics of own side and other side
Perceptions of Conservative supporters Perceptions of Labour supporters
Perceptions of Leavers Perceptions of Remainers
Note: These are mean scores on a 1-5 Likert scale of agreement that these characteristics describe people with a
particular political identity. Data is from the YouGov survey in September 2017. For the party identity
descriptions, the unweighted N is 1,648. For the Brexit identity descriptions, the unweighted N is 1,678.
1 2 3 4 5
Labour ID
Conservative ID
1 2 3 4 5
Labour ID
Conservative ID
1 2 3 4 5
Remainer ID
Leaver ID
1 2 3 4 5
Remainer ID
Leaver ID
In addition to this, when asked about their interest in forms of social interaction with members
of the in-group and out-group, people readily express prejudice toward the out-group and
favouritism toward the in-group. Table 3 shows the proportions of respondents who say that
they would be happy with a child of theirs marrying someone from the other side and the
proportion that are happy to ‘talk politics’ with someone from the other side. Only around half
would be happy to talk politics with the other side, whether that side is defined by Brexit choice
or party identity. Even more strikingly, only a third on average of those with a Brexit identity
would be happy about a prospective son or daughter in-law from the other side. Levels of
partisan prejudice are only slightly higher.
Table 3. Prejudice against the other side
Party identity
Brexit identity
Happy with child marrying other side
YouGov - Sep 2017
Sky – Oct/Nov 2017
Happy to talk politics with other side
YouGov - Sep 2017
Sky – Oct/Nov 2017
Note: The YouGov data have a total unweighted N of 3,326. The Sky data have a total unweighted N of 1,692 for
party identity and 1,702 for Brexit identity
Evaluative bias – perceptions
The final indicator of affective polarization is evaluative bias in perceptions and decision-
making. We start by examining how Brexit identities shape people’s view of the world. There
is a wealth of evidence for the partisan ‘perceptual screen’ when it comes to economic
performance. Supporters of parties in government consistently tend to think that the economy
performed better than supporters of opposition parties (Wlezien et al. 1997; Bartels 2002; De
Boef and Kellstedt 2004; Evans and Pickup 2010; Tilley and Hobolt 2011; Enns 2012; Bisgaard
2015). As Achen and Bartels (2016, 276) bluntly put it, people ‘use their partisanship to
construct “objective facts”’. A similar process of motivated reasoning should apply to people
with Brexit identities. Leavers, who were on the winning side in the referendum, should have
a more positive view of past economic performance than Remainers. We asked respondents in
January 2018 how they thought the economy had performed over the last 12 months on a 1-5
scale (the standard way of measuring retrospective economic perceptions). Table 4 shows the
results of an OLS regression predicting people’s scores on this scale with party identity and
Brexit identity as predictors.13 Higher scores indicate a rosier view of economic performance
during 2017.
Table 4. Predicting retrospective economic perceptions
Brexit identity
No identity
Party identity
Other party
No identity
Adjusted R2
Note: * = p < 0.05. The data come from the January 2018 Tracker survey and have a total unweighted N of 1,418.
The dependent variable asks respondents ‘How do you think the general economic situation in this country has
changed over the last 12 months’ with five options (got a lot worse, got a little worse, stayed the same, got a little
better and got a lot better) coded from 1-5.
13 Table A4a in Appendix 4 shows that including other demographic factors which are correlated with Brexit
identity, such as education and age, makes no difference to these results. Table A4b also shows similar results
using BES data and the more inclusive measure of Brexit identity that is asked of BES respondents. This second
model also includes a measure of British identity which, although correlated with Brexit identity (see table A2c),
has no effect on retrospective economic evaluations.
As expected, there is a gap between Conservative and Labour identifiers in their assessment of
the economy. Conservative identifiers, whose party was in government, were slightly over one-
half of a point on the 1-5 scale more positive about British economic performance in 2017 than
were Labour identifiers. Yet, even holding constant party identity, we see large effects for
Brexit identity. Leavers are almost three-quarters of a point more positive than Remainers. The
effect of Brexit identity is greater than that of party identity in producing biased retrospective
views of the economy.
Evaluative bias – decision making
Another component of evaluative bias that we examine is how decision-making outside the
political realm is shaped by Brexit identities. We are interested in whether these social identities
also spill over into decisions, possibly even discrimination, on non-political matters.
Specifically, we conducted two similar conjoint experiments that asked respondents to choose
between alternative candidates to be Director-General of the BBC and, separately, to be a
lodger in their own home. The advantage of using a conjoint design is that it allows us to
uncover the relative influence of different factors in how people make decisions over bundled
outcomes (Auspurg and Hinz 2014; Hainmueller et al. 2014; Jasso 2006). Borrowed from
marketing research, where it is used to study purchasing decisions, this methodology has
recently been used in public opinion research to study complex opinion formation processes
such as support for immigration policies (Bansak et al. 2016; Hainmueller and Hopkins 2015),
voting for candidates (Hainmueller et al. 2014) and preferences for labour market reform
(Gallego and Marx 2017). In a conjoint study, participants are shown a series of vignettes that
vary according to a determined set of features, with combinations of features randomly varied.
In our studies, each sample was conducted on a distinct sample of approximately 1600
respondents (see Table 1), with each respondent making choices over five pairs of full
randomized candidate profiles. The features in the two designs varied along salient
characteristics, such as age, sex, hobbies, and work experience in the case of the lodger
experiment and age, sex, education and career background for the BBC experiment. In both
experiments we also included two political features: namely, the candidate’s partisan position
in the 2017 UK General Election (Conservative, Labour, or none) and their stance on the 2016
referendum (Leave, Remain, or none).
The full results of preferences for both the Director-General and lodger experiments are in
Appendix 4, but Figures 3 and 4 present the key results. Here we show the marginal mean
outcomes for the two political factors: that is the percentage of times respondents chose profiles
with the specified feature, marginalizing across the other features.14 Figure 3 shows the
marginal means for the party position and referendum position features of the BBC Director-
General experiment separately for people that identify as a Conservative and a Leaver; a
Conservative and a Remainer; Labour and a Leaver; and Labour and a Remainer. There are
large effects of partisanship and Brexit identity. In the upper half of Figure 3, Labour partisans
prefer a Labour supporting Director-General; Conservative partisans prefer a Conservative
supporter. These effects are matched in size by the difference in preferences between Leavers
14 Appendix 4 in the Supplementary Information reports full results in the form of average marginal component
effects (AMCEs; see Hainmueller et al. 2014). Positive AMCEs convey features that make a candidate more
attractive, while negative AMCEs convey features that make a candidate less attractive. The advantages of
marginal means are simplicity of presentation and clarity of base rates for reference categories (Leeper et al.
2019). In general, the factors that we might expect to make for an attractive BBC Director-General (previously
worked as a television producer at the BBC for a long time) and for an attractive lodger (has a job, likes cooking
and does voluntary work) positively affect people’s choices.
and Remainers shown in the lower half of Figure 3. Regardless of someone’s partisanship,
respondents prefer the head of the BBC to have a similar Brexit identity. For example, while
less than 40 per cent of Labour-identified Remainers would pick a candidate who was a Leaver,
holding everything else equal, nearly 60 per cent of Labour-identified Remainers would pick a
fellow Remainer. On average, the effects of Brexit identity are slightly greater than
partisanship. We see very similar patterns in Figure 4 for the lodger experiment. Remainers
prefer to live with a fellow Remainer than a Leaver, and Leavers prefer to live with a fellow
Leaver than a Remainer. Again, these effects are large, and again they are bigger than the
partisan effects. The Brexit divide cross-cuts, and even exceeds, the partisan divide.
Figure 3. Results from BBC Director-General conjoint experiment by Leave and Remain
Note: These are marginal mean outcomes from a discrete choice conjoint experiment, estimated separately for
different types of respondents by their partisan and Brexit identity. Data is from the BBC Director-General
YouGov survey (n=1,653) conducted in October 2017. Error bars reflect 95% confidence intervals, clustered by
respondent with each respondent completing five binary choice decision tasks.
Figure 4. Results from lodger conjoint experiment by Leave and Remain identity
Note: These are marginal mean outcomes from a discrete choice conjoint experiment, estimated separately for
different types of respondents by their partisan and Brexit identity. Data is from the Lodger YouGov survey
(n=1,669) conducted in October 2017. Error bars reflect 95% confidence intervals, clustered by respondent with
each respondent completing five binary choice decision tasks.
‘Political behavior researchers are often struck by the absence of group conflict despite the
existence of distinct and salient groups’ Huddy (2001, 137) has noted. Much research has
therefore focused on the rare cases where long-standing social identities generate considerable
tension, such the partisan divide in the United States or inter-ethnic tensions in other parts of
the world. Yet, we describe a situation in which distinct and salient groups emerged over a
relatively short period of time and generated group conflict on par with that of partisanship.
Building on theories of social identity, we advance the conceptualization of affective
polarization, arguing that such animosity can be mobilized across opinion-based groups in the
context of significant political events. Unlike partisan loyalties, opinion-based groups are
defined by shared opinions on a specific issue or shared cause. We study this phenomenon of
opinion-based group identities in the wake of a critical political juncture: Britain’s 2016
referendum on EU membership. Our results clearly suggest that affective polarization is a
phenomenon not unique to partisanship. Indeed, we show that polarization along the Brexit
divide is as large, or larger, than partisan affective polarization and its effects cross-cut partisan
We thus make a significant contribution to the political behaviour literature by developing the
notion of affective polarization along these opinion-based group lines. Empirically, we
demonstrate these polarization dynamics outside the US context and along nonpartisan lines in
all three areas of affective polarization: identification, differentiation, and evaluative bias.
While theorizing about the origins of affective polarization remains underdeveloped, our work
suggests that long-term ideological polarization, at either the elite or mass level, is unlikely to
be the only cause of new opinion-based identities. Brexit-related identities and polarization
emerged despite no longstanding Leave/ Remain divide and in a manner that cross-cut partisan
identities. This implies that shorter-run dynamics can play an important role in triggering
democratically occurring forms of prejudice, discrimination, and bias. While the empirical
focus in this paper has been Brexit, the notion of affective polarization along opinion-based
group lines could apply elsewhere, where political issues are sufficiently salient and divisive
to give rise to social identities and out-group animosity. For example, this framework could be
applied to the issue of Catalan independence, which has become very politicized and divisive
in Spain, especially in the mobilization leading up to and following the 2017 Catalan
referendum on independence (Criado et al. 2018; Hierro and Gallego 2018; Oller i Sala et al.
At the same time, however, we do not think that all issue debates – regardless of their degree
of underlying disagreement – can generate the consistent and intense patterns of polarization
demonstrated here. Part of the reason for that is the prevalence of the underlying opinion-based
group identities and their perceived importance to large shares of the British public. Although
some people hold views on many different issues and consider those views personally
important, such issue publics are generally understood to be small and narrow (Converse 1964;
Krosnick 1990). In the case of Brexit, opinion-based identities are now held by over three-
quarters of the public and the intensity of those identities is similar to partisanship. The national
referendum and surrounding debate seem necessary, but insufficient, to have generated such
polarization. This is important because not all events of direct democracy, or political debates
more generally, create such deep divides. Referendums are frequent occurrences in many
democracies, yet few appear to generate salient and lasting polarization. In this case, we suspect
that the cross-cutting nature of partisan and Brexit identities plays an important role. Most
national referendums reflect the playing out of elite partisan competition at the mass level
(Prosser 2016) and many EU referendums showcase second-order evaluations of national
governments (Garry et al. 2005; Hobolt 2009). But the Brexit referendum occurred
orthogonally to the traditional partisan divide and has still not been fully subsumed into normal
lines of party competition.
This paper also raises other number of important questions. One such question is how affective
polarization along opinion-based group lines evolves in the long-run: does it fade away as the
political event which triggered the social identities become less salient? It is certainly possible
that Brexit identities will eventually become less important to people now that Britain has left
the EU. Another possibility is that affective polarization on the Brexit issue will lead to a
realignment of the British political system. According to Carmines and Stimson’s seminal work
on issue evolution, realignments are precipitated by the ‘emergence of new issues about which
the electorate has intense feeling that cut across rather than reinforce existing bases of support
for political parties’ (Carmines and Stimson 1981: 107). We have shown that the Brexit
referendum led to the emergence of intensely felt identities that cross-cut partisan divisions.
This could mean that affective polarization along Brexit lines will eventually lead to a more
fundamental change in the UK party system. The major political parties could align their
positions firmly with one of the two opposing positions on future UK-EU relations leading
voters to discard old party attachments in favour of new patterns of support. Indeed, we could
see a similar change in Britain, albeit precipitated in a very different way, to the Southern
realignment in the US (Stanley 1988; Valentino and Sears 2005) and the shift from the main
dimension of party competition being economic left-right policy to social conservative-liberal
policy. Studies of electoral competition in the 2017 and 2019 UK general elections give some
indications that this realignment has already started to occur (Curtice 2017b; Heath and
Goodwin 2017; Jennings and Stoker 2017; Tilley and Evans 2017; Prosser 2018; Hobolt 2018;
Cutts et a. 2020; Hobolt and Rodon 2020).
Whether there is a party realignment or not, it is clear that the EU referendum activated an
important new divide in British society. Intensely felt political division seems to be an all-too-
familiar feature of 21st century democratic politics. Ultimately, any time such division emerges,
normative questions are raised about what this means for democratic society, what might
ameliorate the tension, and how democratic practice might be improved. Answers to these three
questions might be the lack of democratic deliberation, the potential value of a more
deliberative democracy, and deeper institutionalization of deliberative processes, respectively
(Dryzek and Niemeyer 2006; Thompson and Gutmann 1996). The deliberative response is to
seek consensus by an airing of rival arguments. Yet the apparent unwillingness of citizens even
to speak across the divide, let alone respect or befriend one another, would seem to undermine
the possibility of a deliberative cure. Other answers need to be found. The task may not be to
find consensus across the divide, but instead to help citizens to recognize one another not as
enemies and out-groups, but as adversaries with a shared collective identity disagreeing over
the outcomes of policy debate (Mouffe 1999). In that sense, perhaps political scientists, and
political theorists, should move beyond trying to understand how to overcome political
disagreements, and focus more on how those disagreements can be sustained without yielding
deleterious social consequences.
Supplementary material.
Data replication sets are available in Harvard Dataverse at: and online
appendices at:
We are grateful to the Economic and Social Research Council - UK in a Changing Europe
programme for generous support of this research (ES/R000573/1). Previous versions of this
paper have been presented at the University of British Columbia, Durham University, the
University of Exeter, the University of Zurich, the ECPR-SGEU conference in Paris, the EPSA
conference in Vienna and the APSA conference in Boston. We are grateful to the participants
at these events for their insightful comments and suggestions, and in particular we would like
to thank Alexa Bankert, John Curtice, Andy Eggers, Leonie Huddy, Lily Mason, Anand
Menon, Yphtach Lelkes, Toni Rodon and Chris Wratil. We would also like to thank the
anonymous reviewers and the BJPolS editorial team for their constructive assessment and their
guidance. Any remaining errors and imperfections are our own.
Achen, Christopher H. and Larry M. Bartels. 2016. Democracy for Realists: Why elections do
not produce responsive governments. Princeton: Princeton University Press.
Anderson, Christopher J., Andre Blais, Shaun Bowler, Todd Donovan and Ola Listhaug. 2005.
Losers’ Consent: Elections and democratic legitimacy. Oxford University Press.
Auspurg, Katrin, and Thomas Hinz. 2014. Factorial Survey Experiments. Sage Publications.
Bansak, Kirk, Jens Hainmueller, and Dominik Hangartner. 2016. ‘How economic,
humanitarian, and religious concerns shape European attitudes toward asylum seekers.
Science: aag2147.
Bartle, John and Paolo Belluci. 2009. ‘Partisanship, social identity and individual attitudes’, in
Bartle, John and Paolo Belluci, eds., Political Parties and Partisanship: Social identity
and individual attitudes. New York: Routledge.
Bartels, Larry M. 2002. ‘Beyond the Running Tally: Partisan Bias in Political Perceptions.’
Political Behavior 24(2): 117-150.
Becker, Sascha O., Thiemo Fetzer, and Dennis Novy. 2017. ‘Who voted for Brexit? A
comprehensive district-level analysis.’ Economic Policy 32(92): 601-650.
Bisgaard, Martin. 2015. ‘Bias will find a way: Economic perceptions, attributions of blame,
and partisan-motivated reasoning during crisis.’ Journal of Politics 77(3): 849-860.
Bliuc, Ana-Maria, Craig McGarty, Katherine Reynolds, and Daniela Muntele. 2007. ‘Opinion-
based group membership as a predictor of commitment to political action.’ European
Journal of Social Psychology 37(1): 19-32.
Cain, Bruce E., and John Ferejohn. 1981. "Party Identification in the United States and Great
Britain." Comparative Political Studies 14.1: 31-47.
Carmines, Edward G., and James A. Stimson. 1981. ‘Issue evolution, population replacement,
and normal partisan change.’ American Political Science Review 75(1): 107-118.
Clarke, Harold D., Matthew Goodwin, and Paul Whiteley. 2017. Brexit: Why Britain voted to
leave the European Union. Cambridge, UK: Cambridge University Press.
Colantone, Italo, and Piero Stanig. 2018. "Global competition and Brexit." American Political
Science Review 112.2: 201-218.
Converse, Philip. 1964. ‘The nature of belief systems in mass publics’, in Apter, David, ed.
Ideology and Discontent. New York: Free Press.
Criado, Henar, Francisco Herreros, Luis Miller, and Paloma Ubeda. 2018. The unintended
consequences of political mobilization on trust: The case of the secessionist process in
Catalonia.’ Journal of Conflict Resolution 62 (2): 231-253.
Curtice, John. 2017a. ‘Why Leave won the UK’s EU referendum.Journal of Common Market
Studies 55(S1): 19-37.
Curtice, John. 2017b. ‘General Election 2017: A new two party politics?’ Political Insight 8(2):
Cutts, David, Matthew Goodwin, Oliver Heath, Paula Surridge (2020), ‘Brexit, the 2019
General Election and the realignment of British politics’, Political Quarterly
Dahl, Robert A. 1967. Pluralist Democracy in the United States: Conflict and consent. New
YorK: Rand McNally.
De Boef, Suzanna, and Paul M. Kellstedt. 2004. ‘The political (and economic) origins of
consumer confidence.’ American Journal of Political Science 48: 633–49.
Dryzek, John S. and Simon Niemeyer. 2006. ‘Reconciling pluralism and consensus as political
ideals.’ American Journal of Political Science 50(3): 634-649.
Electoral Commission. 2015. ‘Referendum on membership of the European Union:
Assessment of the Electoral Commission on the proposed referendum question.’ Report.
Enns, Peter K., Paul M. Kellstedt and Gregory E. McAvoy. 2012, ‘The consequences of
partisanship in economic perceptions.’ Public Opinion Quarterly 76(2): 287-310
Evans, Geoffrey, and James Tilley. 2017. The New Politics of Class: The political exclusion of
the British working class. Oxford, UK: Oxford University Press.
Evans, Geoffrey, and Anand Menon. 2017. Brexit and British Politics. London: Polity.
Evans, Geoffrey, and Mark Pickup. 2010. ‘Reversing the causal arrow: The political
conditioning of economic perceptions in the 2000-2004 U.S. Presidential cycle.’ Journal
of Politics 72(4): 1236-51.
Fieldhouse, Edward, Jane Green, Geoffrey Evans, Herman Schmitt, Cees van der Eijk, Jon
Mellon and Chris Prosser. 2019. British Election Study 2014-2024 Combined Waves 1-15
Internet Panel.
Ford, Robert, and Matthew Goodwin. 2017. ‘Britain after Brexit: A nation divided.Journal
of Democracy 28: 17-30.
Fiorina, Morris P., and Samuel J. Abrams. 2008. ‘Political polarization in the American public.’
Annual Review of Political Science 11: 563-588.
Gallego, Aina, and Paul Marx. 2017. ‘Multi-dimensional preferences for labour market
reforms: a conjoint experiment.’ Journal of European Public Policy, 24(7): 1027-1047
Garry, John, Michael Marsh, and Richard Sinnott. 2005. ‘“Second-order” versus “Issue-
voting” Effects in EU Referendums: evidence from the Irish nice treaty referendums.’
European Union Politics 6(2): 201-221.
Goodwin, Matthew J., and Oliver Heath. 2016. ‘The 2016 referendum, Brexit and the left
behind: an aggregate-level analysis of the result.’ Political Quarterly 87(3): 323-332.
Green, Donald P., Bradley Palmquist, and Eric Schickler. 2004 Partisan Hearts and Minds:
Political parties and the social identities of voters. Yale University Press.
Greene, Steven. 2000. ‘The psychological sources of partisan-leaning independence.
American Politics Quarterly 28(4): 511-537.
Hainmueller, Jens and Daniel J. Hopkins. 2015. ‘The hidden American immigration consensus:
A conjoint analysis of attitudes toward immigrants.’ American Journal of Political Science
59: 529–48.
Hainmueller, Jens, Daniel J. Hopkins, and Teppei Yamamoto. 2014. ‘Causal inference in
conjoint analysis: understanding multidimensional choices via stated preference
experiments.’ Political Analysis 22(1): 1-30.
Heath, Oliver, and Matthew Goodwin. 2017. ‘The 2017 General Election, Brexit and the return
to two-party politics: An aggregate-level analysis of the result.Political Quarterly 88(3):
Hetherington, Marc J. 2009. ‘Putting polarization in perspective.’ British Journal of Political
Science 39(2): 413-448
Hierro, María José, and Aina Gallego. 2018. ‘Identities in between: Political conflict and
ethnonational identities in multicultural states.’ Journal of Conflict Resolution 62 (6):
Hobolt, Sara B. 2018. ‘Brexit and the 2017 UK General Election.’ Journal of Common Market
Studies 56, 39-50.
Hobolt, Sara B. 2016. ‘The Brexit vote: a divided nation, a divided continent.’ Journal of
European Public Policy 23(9): 1259-1277.
Hobolt, Sara B. 2009. Europe in Question. Referendums on European Integration. Oxford,
UK: Oxford University Press.
Hobolt, Sara B., and Toni Rodon. 2020. ‘Cross-cutting issues and electoral choice. EU issue
voting in the aftermath of the Brexit referendum.Journal of European Public Policy 27
(2): 227-245.
Horowitz, Donald L. 1993. ‘Democracy in divided societies.’ Journal of Democracy 4(4): 18-
Hooghe, Liesbet, and Gary Marks. 2018. ‘Cleavage theory meets Europe’s crises: Lipset,
Rokkan, and the transnational cleavage.Journal of European Public Policy 25(1): 109-
Huddy, Leonie. 2001. ‘From social to political identity: A critical examination of social identity
theory.’ Political Psychology 22(1): 127-156.
Huddy, Leonie, Lilliana Mason, and Lene Aarøe. 2015. ‘Expressive partisanship: Campaign
involvement, political emotion, and partisan identity.’ American Political Science Review,
Ipsos-Mori. 2018. ‘Issues Index: 2007 onwards.’ Accessed 4 June 2018 from
Iyengar, Shanto, Guarav Sood, and Yphtach Lelkes. 2012. ‘Affect, Not Ideology: A Social
Identity Perspective on Polarization.’ Public Opinion Quarterly 76(3): 405-431.
Iyengar, Shanto, and Sean J. Westwood. 2015. ‘Fear and loathing across party lines: New
evidence on group polarization.’ American Journal of Political Science, 59(3): 690-707.
Iyengar, Shanto, Yphtach Lelkes, Matthew Levendusky, Neil Malhotra, and Sean J. Westwood.
2019. "The origins and consequences of affective polarization in the United States."
Annual Review of Political Science 22: 129-146.
Jasso, Guillermina. 2006. ‘Factorial survey methods for studying beliefs and judgements.
Sociological Methods & Research 34(3): 334-423
Jennings, Will, and Gerry Stoker. 2017. ‘Tilting towards the cosmopolitan axis? Political
change in England and the 2017 General Election.Political Quarterly 88(3): 359-369.
Johnston, Ron, D. Manley, K. Jones, and Charles J. Pattie. 2018. ‘Geographies of Brexit and
its aftermath: Voting in England at the 2016 referendum and the 2017 General Election.
Space and Polity: In press.
Kinder, Donald R., and Nathan P. Kalmoe. 2017. Neither Liberal nor Conservative:
Ideological innocence in the American public. Chicago: University of Chicago Press.
Klar, Samara. 2014. ‘Partisanship in a social setting.’ American Journal of Political Science
58.3 (2014): 687-704.
Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier, and
Timotheos Frey. 2008. West European Politics in the Age of Globalization. Cambridge,
UK: Cambridge University Press.
Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier, and
Timotheos Frey. 2006. ‘Globalization and the transformation of the national political
space: Six European countries compared.’ European Journal of Political Research 45(6): 921-956.
Krosnick, Jon C. 1990. ‘Government policy and citizen passion: A study of issue publics in
contemporary America.’ Political Behavior 12(1): 59-92.
Lacombe, Matthew J. 2019. "The Political Weaponization of Gun Owners: the National Rifle
Association's Cultivation, Dissemination, and Use of a Group Social Identity." The
Journal of Politics 81(4): 1342-1356.
Layman, Geoff C., Thomas M. Carsey, and Juliana Menasce Horowitz. 2006. ‘Party
polarization in American politics: Characteristics, causes, and consequences.’ Annual
Political Science Review 9: 83-110.
Leeper, Thomas, Sara Hobolt and James Tilley. 2019. ‘Measuring subgroup preferences in
conjoint experiments’, Political Analysis, DOI:10.1017/pan.2019.30
Lelkes, Yphtach, and Sean J. Westwood. 2017. ‘The limits of partisan prejudice.’ Journal of
Politics 79(2): 485-501.
Levendusky, Matthew S. 2013. ‘Why do partisan media polarize viewers?’ American Journal
of Political Science 57(3): 611-623.
Levendusky, Matthew and Neil Malhotra. 2016. ‘Does media coverage of partisan polarization
affect political attitudes?’ Political Communication 33(2): 283-301.
Lipset, Seymour Martin. 1959. Political Man: The social bases of politics. Baltimore: Johns
Hopkins University Press.
MacKuen, Michael, Jennifer Wolak, Luke Keele, and George E. Marcus. 2010. ‘Civic
engagements: Resolute partisanship or reflective deliberation.’ American Journal of
Political Science 54(2): 440-458.
Malka, Ariel, and Yphtach Lelkes. 2010. ‘More than ideology: Conservative–liberal identity
and receptivity to political cues.’ Social Justice Research 23(2-3): 156-188.
Mason, Lilliana. 2013. ‘The rise of uncivil agreement: Issue versus behavioral polarization in
the American electorate.’ American Behavioral Scientist 57(1): 140-159.
Mason, Lilliana. 2015. ‘“I disrespectfully agree”: The differential effects of partisan sorting on
social and issue polarization.’ American Journal of Political Science 59(1): 128-145.
Mason, Lilliana. 2018. Uncivil Agreement: How politics became our identity. Chicago:
University of Chicago Press.
McGarty, Craig, Ana-Maria Bliuc, Emma F. Thomas, and Renata Bongiorno. 2009. ‘Collective
action as the material expression of opinion-based group membership.’ Journal of Social
Issues 65(4): 839-857.
Miller, Patrick R., and Pamela Johnston Conover. 2015. ‘Red and blue states of mind: Partisan
hostility and voting in the United States.’ Political Research Quarterly 68(2): 225-239.
Mouffe, Chantal. 1999. ‘Deliberative Democracy or Agonistic Pluralism?’ Social Research
66(3): 745-748.
Oller i Sala, Josep Maria, Albert Satorra, and Adolf Tobeña. 2019. ‘Secessionists vs. unionists
in Catalonia: mood, emotional profiles and beliefs about secession perspectives in two
confronted communities.’ Psychology 10(3): 336-357.
Prosser, Christopher. 2016. ‘Calling European Union treaty referendums: electoral and
institutional politics.’ Political Studies 64(1): 182-199.
Prosser, Christopher. 2018. ‘The strange death of multi-party Britain: The UK General Election
of 2017.’ West European Politics 41(5): 1226-1236.
Rabushka, Alvin, and Kenneth A. Shepsle. 1972. Politics in Plural Societies. Redwood City,
CA: Stanford University Press.
Richards, Lindsay, and Anthony Heath. 2017. CSI Brexit 1: How much are people willing to
pay for the Brexit divorce bill? Centre for Social Investigation.
Schattschneider, Elmer. 1960. The Semi-Sovereign People. New York: Holt, Rinehart and
Smith, Laura G.E., Emma F. Thomas, and Craig McGarty. 2015. ‘“We must be the change we
want to see in the world”: Integrating norms and identities through social interaction.’
Political Psychology 36(5): 543-557.
Stanley, Harold W. 1988. ‘Southern partisan changes: Dealignment, realignment or both?’
Journal of Politics 50(1): 64-88.
Tajfel, Henri. 1970. ‘Experiments in intergroup discrimination.’ Scientific American 223(5):
Tajfel, Henri. 1979. ‘Individuals and groups in social psychology.’ British Journal of Clinical
Psychology 18(2): 183-190.
Tajfel, Henri. 1982. ‘Social psychology of intergroup relations.’ Annual Review of Psychology
33(1): 1-39.
Tajfel, Henri, and John C. Turner. 1979. ‘An integrative theory of intergroup conflict’ p33-47
in Austin, William G. and Stephen Worchel, eds., The Social Psychology of Intergroup
Relations. Monterey, CA: Brooks/Cole.
Thompson, Dennis, and Amy Gutmann. 1996. Democracy and Disagreement. New York:
Cambridge University Press.
Tilley, James, and Sara B. Hobolt. 2011. ‘Is the government to blame? An experimental test of
how partisanship shapes perceptions of performance and responsibility.’ Journal of
Politics 73(2): 316-330.
Tilley, James, and Geoffrey Evans. 2017. ‘The New Politics of Class after the 2017 General
Election.’ Political Quarterly 88(4): 710-715.
Turner, John C. 1982. ‘Towards a cognitive redefinition of the social group.’ In Tajfel, Henri,
ed., Social Identity and Intergroup Relations. Cambridge: Cambridge University Press, 15-
Valentino, Nicholas A., Vincent L. Hutchings, Antoine J. Banks, and Anne K. Davis. 2008. ‘Is
a worried citizen a good citizen? Emotions, political information seeking, and learning via
the internet.’ Political Psychology 29(2): 247-273.
Valentino, Nicholas A., and David O. Sears. 2005. ‘Old times there are not forgotten: Race and
partisan realignment in the contemporary South.’ American Journal of Political Science
49(3): 672-688.
Westwood, Sean J., Shanto Iyengar, Stefaan Walgrave, Rafael Leonisio, Luis Miller, and
Oliver Strijbis. 2018. ‘The tie that divides: Cross-national evidence of the primacy of
partyism.’ European Journal of Political Research 57 (2): 333-354.
Wlezien, Christopher, Mark Franklin, and Daniel Twiggs. 1997. ‘Economic perceptions and
vote choice: Disentangling the endogeneity.’ Political Behavior 19(1): 7-17.
... The referendum about membership of the UK within the European Union was accompanied by different debates on "national sovereignty" and "independence". These debates took place within different discursive contexts and they were characterised by various argumentative strategies, political calculations and public emotions (Davies, 2021;Hobolt, Leeper, Tilley, 2020;Richards, Heath, Elgenius, 2020). Flemmen & Savage (2017) show that the Brexit referendum situated itself within a long history of nationalist sentiment among the "white" population of the UK. ...
This paper analyses the transformation of identities and institutions of economic experts as politico-economic professionals brought about by the Brexit process between 2016 and 2020. In order to understand the transformation from order to chaos in times of COVID-19, a mixed methods approach is applied. In a first step, a discourse analysis of two competing letters from two different groups of economists is carried out. One letter was published by “Economists for Remain” before the Brexit referendum in 2016; the other letter was published by a group called “Economists for Free Trade”, arguing for leaving the EU. In a second step, we conduct a field and social space analysis of economists who signed the letters in order to analyse their social backgrounds and professional positions within the UK academic, media, political and economic system. The paper shows that members of both groups take significantly different positions. In a final step, we conducted interviews in 2019/20 with selected members of each group. Here we show how the professional system of UK economic governmentality enters a state of crisis. The paper finally argues that contemporary nationalist movements cannot provide a new social order. They lead societies into chaos instead.
... 4.2. Dependent variable: Affective polarization at the regional level Affective polarization has been measured in various ways, using feeling thermometer ratings (Iyengar et al., 2019;Wagner, 2020), traits ratings (Iyengar et al., 2012;Levendusky, 2018), trust (Westwood et al., 2018), opinion-based groups (Hobolt et al., 2021) or social distance measures; albeit the latter has been characterized as less consistent (Druckman & Levendusky, 2019). Thermometer ratings, using survey data asking respondents to evaluate parties or partisans on a 0-100 scale, are the most common tools used in the literature. ...
Full-text available
This paper investigates divergences in levels of affective polarization across Belgian regions around the 2019 elections. Elaborating on the relative deprivation theory, we analyse the role of current and long-term socioeconomic regional inequalities and of geographical distance separating regions. Empirically, we aggregate individual-level measures of affective polarization at the Belgian NUTS-3 level and use a gravity approach to explore the determinants of regional divergences. Our results show that regional variations in affective polarization are best explained by a current rural-urban divide and by how regions have performed economically in the last years. We also show that geographical proximity matters and reinforces the effects of economic deprivation.
... For instance, the cases of Spain and the United Kingdom show that no analysis of trust in the European Parliament can abstract from the context of their political realities. In the UK, the debate over Brexit, i.e., the ultimate abandonment of EU membership, proved among the most polarising political topic of the decade (Hobolt et al. 2021). Unlike in most other EU member states, this translated into substantial and continuous public sphere attention devoted to European institutions (North et al. 2021). ...
Full-text available
Differences in the response-scale formats constitute a major challenge for ex-post harmonisation of survey data. Linear stretching of original response options onto a common range of values remains a popular response to format differences. Unlike its more sophisticated alternative, simple stretching proves readily applicable without requiring assumptions regarding scale length or access to auxiliary information. The transformation only accounts for response scale length, ignoring all other aspects of measurement quality, which makes the equivalence of harmonised survey variables questionable. This paper focuses on the inherent limitations of linear stretching based on a case study focusing on the measurements of trust in the European Parliament by the Eurobarometer and the European Social Survey—8 timewise corresponding survey waves in 14 European countries (2004–2018). Our analysis demonstrates that the linear stretch approach to harmonising question items with different underlying response scale formats does not make the results of the two surveys equivalent. Despite harmonisation, response scale effects are retained in the distributions of output variables.
... The merging of economics and culture also is evident in Green et al., (2022), who argue that economic concerns influenced Brexit views perceived through the lens of relative gains and losses of various social groups. Hobolt et al., (2021) demonstrate the extent to which the referendum solidified identities and intensified polarization. ...
Full-text available
In this paper we examine three forms of regret in relation to the UK’s hugely significant referendum on EU membership that was held in June 2016. They are: (i) whether ‘leave’ voters at the referendum subsequently regretted their choice (in the light of the result), (ii) whether non-voters regretted their decisions to abstain (essentially supporting ‘remain’) and (iii) whether individuals were more likely to indicate that it is everyone’s duty to vote following the referendum. We find evidence in favor of all three types of regret. In particular, leave voters and non-voters were significantly more likely to indicate that they would vote to remain given a chance to do so again; moreover, the probability of an individual stating that it was everyone’s duty to vote in a general election increased significantly in 2017 (compared to 2015). The implications of the findings are discussed in the context of the referendum’s outcome.
Public opinion polarisation can impair society’s ability to reach a democratic consensus in different political issue areas. This appears particularly true when the polarisation of opinions coincides with clearly identifiable social groups. The literature on public opinion polarisation has mostly focussed on the US two-party context. We lack concepts and measures that can be adapted to European countries with multi-party systems and multi-layered group identities. This article proposes a conceptualisation of polarisation between groups in society. It presents a measure that captures the overlap of ideology distributions between groups. The two-step empirical framework includes hierarchical IRT models and a measure for dissimilarity of distributions. The second part presents an empirical application of the measure based on survey data from Switzerland (1994–2016), which reveals insightful dynamics of public opinion polarisation between party supporters and education groups.
Political contexts affect the degree to which traditional campaign methods like door knocking and leafleting affect referendum outcomes. Campaign methods are most effective when issues are of low salience and political parties provide unified cues to their supporters. In this paper, I draw on Zaller’s model of opinion formation to derive some theoretical expectations about the effectiveness of campaign contact in referendum campaigns when the issue is highly salient but the political parties provide mixed cues, arguing that prior voting intention, rather than partisan identity, should affect receptiveness to campaign messages. I test these expectations with regression analysis of British Election Study panel data, using propensity score matching to mitigate selection bias. My results suggest that the effects of campaigning during the Brexit referendum were mostly small, but differed significantly according to an individual’s political attention and, contrary to the theoretical expectations, their partisan identity.
Full-text available
We analyze how parties respond programmatically to populist parties in power abroad. Political parties often copy the policies of governing parties in other countries–a phenomenon that contributes to waves of transnational policy diffusion. We report the first large-scale comparative study showing that populist parties in government abroad trigger the opposite reaction: instead of inspiring emulation, their highly visible governing dilemmas provoke a policy backlash by parties in other states. We argue that dilemmas arise because populist parties confront unique and debilitating trade-offs between maintaining their anti-system posture and governing effectively, which make them electorally vulnerable. Other parties observe foreign populists’ governing dilemmas and respond by distancing themselves in order to avoid these problems. We detect this “foreign populist backlash effect” using spatial econometric analysis, a method that allows us to estimate international policy connections between parties, applied to over 200 European parties’ programmatic positions since the 1970s. Our findings illuminate parties’ election strategies and show that this backlash effect constrains the spread of populism across Western democracies.
Computer mediated communication has marked differences from the face-to-face context. One major difference is that, in the online context, we often have explicit access to others' opinions and these opinions are often the only informational cues available. We investigate if awareness of opinion congruence, in the absence of any other reference categories, may be sufficient to foster social identification. In a pre-registered experiment (N = 681), we manipulated exposure to opinions, and measured levels of ingroup identification, opinion-based identification and their social influence on activism intentions. Our results demonstrate exposure to others' opinions in an otherwise anonymous context fosters ingroup and opinion-based identification. There was no effect on opinion-based group activism intentions. We conclude that computer mediated contexts have consequences for identification – opinion (in)congruence is becoming more relevant as a source of social categorization. While we did not find this identification had a social influence on activism, we discuss avenues for future research to disentangle the features of opinion-based groups necessary to foster activism.
The Trade and Cooperation Agreement (TCA) is the start of a new relationship between the UK and the European Union (EU). As the consequences of Brexit unfold, there will be pressure to change the TCA, either in a “softer” or “harder” direction. To determine the potential medium-term direction of the EU–UK relationship, we conducted a conjoint survey experiment with a sample of British voters, where we asked them to choose between different hypothetical package deals. When faced with such choices, British citizens overall mostly support a softer relationship in which the UK applies EU regulatory standards in return for greater access to the single market. However, Leave voters most often support a much harder trade-off of full regulatory sovereignty but continued restrictions on UK exports.
Full-text available
There is a continuing debate over the political importance and durability of partisan attachments in European multi-party systems. Drawing on a nationally representative 5-wave panel, we provide a longitudinal test of the power of partisanship in Italy over the course of the tumultuous 2013 national elections. We find that a strong partisan affiliation measured as a social identity two years prior to the election promoted system stability by increasing support for the in-party and inhibiting electoral support for the insurgent Five Star Movement (M5S). In contrast, non-partisans, especially highly educated ones, were more likely than partisans to vote for M5S. Our results illustrate the role of partisanship in stabilizing multi-party systems amid crisis.
The outcome of the 2019 general election—a resounding Conservative majority and an unprecedented defeat for Labour—delivered a decisive electoral verdict for the first time in recent years following a period where British politics has been characterised by instability and indecision. In this article, we draw on aggregate‐level data to conduct an initial exploration of the vote. What was the impact of Brexit on the 2019 general election result? How far has Brexit reshaped electoral politics? Was 2019 a ‘realignment election’? And, if so, what are the implications? With a focus on England and Wales we show that, although the Conservatives made gains deep into Labour’s working class heartlands, these gains have been a long time coming, reflected in Labour’s weakening relationship with working class Britain. As such, 2019 is not a critical election but a continuation of longer‐term trends of dealignment and realignment in British politics.
As political competition is becoming increasingly multi-dimensional in Europe, voters often face the challenge of choosing which issues matter most to them. The European integration issue presents a particular difficulty for voters, since it is not closely aligned to the left-right dimension. We test the impact of the EU issue in the first parliamentary election following the UK's divisive Brexit referendum. We argue that while the EU issue was salient to voters, EU issue voting was inhibited by the indistinct and ambiguous positions adopted by the two major parties. To examine this, we combine an analysis of British Election Study data from the 2017 General Election with a conjoint experiment that allows us to present voters with a range of choices on both dimensions. Our findings show that the EU dimension has the potential to become a cross-cutting dimension that rivals the left-right dimension in British electoral politics, but this crucially depends on party competition.
Conjoint analysis is a common tool for studying political preferences. The method disentangles patterns in respondents’ favorability toward complex, multidimensional objects, such as candidates or policies. Most conjoints rely upon a fully randomized design to generate average marginal component effects (AMCEs). They measure the degree to which a given value of a conjoint profile feature increases, or decreases, respondents’ support for the overall profile relative to a baseline, averaging across all respondents and other features. While the AMCE has a clear causal interpretation (about the effect of features), most published conjoint analyses also use AMCEs to describe levels of favorability. This often means comparing AMCEs among respondent subgroups. We show that using conditional AMCEs to describe the degree of subgroup agreement can be misleading as regression interactions are sensitive to the reference category used in the analysis. This leads to inferences about subgroup differences in preferences that have arbitrary sign, size, and significance. We demonstrate the problem using examples drawn from published articles and provide suggestions for improved reporting and interpretation using marginal means and an omnibus F-test. Given the accelerating use of these designs in political science, we offer advice for best practice in analysis and presentation of results.
While previously polarization was primarily seen only in issue-based terms, a new type of division has emerged in the mass public in recent years: Ordinary Americans increasingly dislike and distrust those from the other party. Democrats and Republicans both say that the other party’s members are hypocritical, selfish, and closed-minded, and they are unwilling to socialize across party lines. This phenomenon of animosity between the parties is known as affective polarization. We trace its origins to the power of partisanship as a social identity, and explain the factors that intensify partisan animus. We also explore the consequences of affective polarization, highlighting how partisan affect influences attitudes and behaviors well outside the political sphere. Finally, we discuss strategies that might mitigate partisan discord and conclude with suggestions for future work. Expected final online publication date for the Annual Review of Political Science Volume 22 is May 11, 2019. Please see for revised estimates.
Much has been written since the 2016 Brexit referendum regarding the divides within British society that the vote illustrated – including geographical divides – and their influence on the outcome of the 2017 general election. Focusing on England, this paper explores the extent and significance of those geographical divides at the 2016 referendum, at a variety of spatial scales – concluding that apart from a major difference between parts of inner London and the rest of England these were largely insignificant. Turning to the 2017 general election, analyses show that this return to a predominantly two-party system within England largely involved a replication of the geography of the 2015 general election outcome. A new electoral map of England did not emerge from the divisions that Brexit stimulated: the country is divided along class lines, with London standing out as different from all other regions.