Voter Decision-Making in a Context of Low Political Trust: The 2016 UK EU Membership Referendum

Abstract

Using volunteer writing for Mass Observation, we explore how British citizens decided whether to leave the EU. The 2016 referendum was the biggest decision made by the British electorate in decades, but involved limited voter analysis. Many citizens did not have strong views about EU membership in early 2016. The campaigns did not help to firm up their views, not least because so much information appeared to be in dispute. Voters, often characterised as polarised, were reluctant and uncertain. Many citizens took their duty to decide seriously, but were driven more by hunch than careful analysis. In 2016, voters reacted against elites they did not trust at least as much as they embraced the ideas of trusted elites. This contrasts with the 1975 Referendum on the Common Market, when the vote was driven by elite endorsement. In low-trust contexts, voters use cues from elites as negative rather than positive stimulus.

Introduction

In the period since the EU Referendum of 2016, we have learned a lot about what people thought of the EU, Brexit, and related issues, and how this varied by locality, social group, and attitude to other issues (Clarke et al., 2017a, 2017b; Curtice, 2016; Goodwin and Heath, 2016; Goodwin and Milazzo, 2017; Hobolt, 2016; Lee et al., 2018). However, we have learned rather less about how people thought of these things, how they formed opinions in response to requests from politicians and pollsters, and how they came to decisions as voters in the period leading up to the referendum. This article aims to fill this gap using evidence collected by the Mass Observation (MO) Project.

We begin by emphasising the unique contribution of MO collections. The letters and diaries collected from panellists over a period of months in the run-up to the vote allow researchers to explore how opinion and decision-making develops over time. Other qualitative methods such as focus groups allow us to explore the immediate thinking of participants, but MO data make visible the development of thinking.

The results of our study make up the bulk of the article. The MO panellists behave in a way that is in tune with much public opinion research largely drawn from survey work. They are rather vague and uncertain in their deliberations about the issues at stake. They often develop opinions reluctantly, hesitantly, uncertainly, out of a sense of duty as citizens, and in response to surveys, ballots and associated campaigns. In developing these opinions, campaigns often do not help when they are characterised by claim and counter-claim, and take place in a context of low political trust. In developing opinions, therefore, the MO panellists often fall back on feelings and elite cues. Indeed, our evidence supports the argument that voting in the EU Referendum was less an expression of polarisation and more an awkward journey. Our findings encourage caution regarding claims that polarisation produced the conditions for the referendum and the closely divided result (Ford and Goodwin, 2017; Goodhart, 2017; Norris and Inglehart, 2019). They support the counter argument that polarisation was produced by the referendum and associated campaign, outcome and aftermath (Curtice, 2018; Hobolt et al., 2020) – and may fade unless restimulated.

Another main finding of the research is perhaps more surprising. MO collections provide evidence that voters were more uncertain than could be revealed by the forced choices presented in surveys. Many voters found that campaigns and the information environment made their task harder, rather than performing the standard role of clarifying choices and issues. Voters ended up using cues, as much previous research would have predicted, but often in a way not expected. Rather than elite cues steering voters, many citizens voted against the elites they least trusted and so used their distrust of politics to steer their decision-making.

By analysing MO’s referendum diaries, we aim to advance understandings of Brexit, including what explains it and what might be its consequences. Primarily, though, we aim to advance understandings of how people form opinions and make decisions as voters, especially in the current period of low political trust in Britain and many other democracies (Clarke et al., 2018; Stoker, 2016). Elite cues have been a particular focus of research on referendum voting in recent years, not least because referendums often ask citizens to decide on complex and unfamiliar issues (Hobolt, 2006; LeDuc, 2002; Nemčok et al., 2019). Both elite cues and feelings have been much discussed in research on the Brexit vote – and before that, on Euroscepticism across Europe – where the most influential framework has probably been the ‘calculation, community, and cues’ framework of Hooghe and Marks (2005). Here, ‘calculation’ refers to the cost–benefit analysis of slow-thinking and rational choice models, while ‘community’ refers to feelings of national belonging and understandings of national identity (which function as fast-thinking heuristics), and ‘cues’ refers to the fast-thinking heuristic of elite endorsements. Since the EU Referendum of 2016, this framework has been used by Hobolt (2016) and Clarke et al. (2017b), who found the Leave vote to be explained by a combination of all three factors. In the sections that follow, we use MO sources to demonstrate how a context of low political trust in 2016 made calculation difficult for many voters in the referendum (section ‘Unhelpful Campaigns’). These voters ended up basing their decisions more on vague feelings regarding beliefs, principles and visions (similar to the ‘community’ of Hooghe and Marks), and cues – but negative cues, as opposed to the elite endorsements often focused on by studies of cue-taking (section ‘Recourse to Feelings and Elite Cues’).

King (1977) argued that elite cues were particularly important in shaping the result of the UK’s 1975 Referendum on the Common Market. During the 1960s and 1970s, Europe was not a very salient issue for voters. Opinions towards the Common Market fluctuated over time, which King took as an indication that such opinions were lightly held by citizens. Once the two main parties finally decided to support remaining in the Common Market, the majority of citizens fell in behind this position. King referred to this phenomenon as ‘follow the leader’. He noted that little ‘anti-establishment mood’ was present in Britain at the time. Voters mostly followed the endorsements of the main party leaders and also business leaders. In the rest of this article, we provide evidence that something different happened in 2016 when the vote took place against a background of significant anti-establishment mood in Britain.

Approaching Mass Observation

MO was established in 1937 to record the everyday lives of ordinary people in Britain (Hinton, 2013). Initially, most of its commissions came from the Ministry of Information. Until 1965, it collected material by two general means. A team of ‘mass observers’ recorded observations, overhead conversations, survey responses, interview responses and ephemera. A panel of volunteer writers, between 400 and 1000 strong (depending on the year), kept monthly diaries, completed day surveys and replied to quarterly open-ended questions (called ‘directives’). In 1969, a deal was struck with the University of Sussex to archive the papers of this original MO. The Mass Observation Archive (MOA) was opened in 1975. In 1981, the Archive founded the MO Project, which revived the panel of volunteer writers. To this day, directives are still being sent 3 times a year to approximately 500 respondents.

Compared to focus-group talk, often used to explore voter decision-making from a qualitative perspective, writing for MO is done not in public but in private. It is done not in the relatively structured context of one brief meeting but as part of a relatively long-term, unstructured, intimate correspondence between MO and its panellists. In this context, panellists seem prepared to write things for MO that might not get said, or said so frankly, in focus groups or even one-to-one interviews – for example, admissions regarding attitudes to faith or sex (Hinton, 2013; Sheridan, 1994). Furthermore, panellists often use the format of a letter when writing for MO, which affords more time and space for reflection and to include more detail than is often provided by other formats. Perhaps most importantly for the rest of this article, panellists sometimes use the format of a diary when writing for MO, especially around the time of elections or referendums. These diaries allow researchers to reconstruct what happens during political campaigns. They provide researchers with access to how citizens receive requests for opinions and votes, and how they respond – sometimes right up to the point of casting their vote.

The observations of the original mass observers have been criticised for telling historians more about the prejudices of these untrained ethnographers than the opinions, values and understandings of ordinary citizens (MacClancy, 1995). The volunteer writing of MO panellists, as ‘the most unmediated layer’ of the archive (Sheridan, 1994), is generally thought to be less open to such criticism. In what follows, we limit our focus to this volunteer writing. We selected two directives on Brexit from 2016 and 2017 (see Appendix 1). Panellists took between 1 and 4 months to respond in each case. The average length of responses was nine sides of A4. A few panellists limited their focus to the particular questions asked in the directives, but many panellists used the title of the directive – ‘the EU Referendum’ – as a prompt to write with apparent freedom about the general topic (a writing practice common to many MO panellists and directives – Sheridan et al., 2000). Connected to this, while some panellists chose a letter format for their responses and sent relatively short responses (less than five sides, apparently written in one sitting), others chose a diary format, sending tens of sides in response (apparently written in multiple instalments over many weeks).

We sampled 60 panellists for each directive, filling quotas for age group, gender, occupational classification, region and Leave/Remain. The figure of 60 was enough to reach descriptive saturation – the most important consideration when sampling qualitative data (Baker and Edwards, 2012). The characteristics of the panellists quoted in the sections below are summarised in Appendix 2.

The social constitution of the original panel (1930s–1960s) has been criticised for being skewed towards the radicalised lower middle class (Jeffery, 1978) or at least people from London and Southeast England, and people on the Left (Hinton, 2013). However, findings generated from the original panel have been validated using findings from Gallup polls and the Wartime Social Survey (Goot, 2008) and also the Home Intelligence Panel (Hinton, 2013) – so these concerns should not be overplayed. Furthermore, the panel constituted since 1981 is widely recognised to be more representative of the British population than the original panel, though it remains skewed a little towards retired and middle-class people, and especially those people who volunteer for a social history project, and so are particularly dutiful, engaged, reflexive and critical (Hinton, 2010).

This presence of especially engaged citizens is not particularly unusual in public opinion research. For example, the Inquiry into the 2015 British General Election Opinion Polls (Sturgis et al., 2016) found people with greater-than-average interest in politics to be over-represented generally in opinion polls. On one hand, then, issues of representativeness should not be overemphasised with regard to MO. The current panel is diverse and contains a range of respondents writing from a range of positions. On the other hand, given the remaining limitations of the panel, how the writing of panellists is approached deserves further consideration. MO sources are most commonly analysed by what might be called ‘reading vertically’ for autobiographical life histories. This is what Dorothy Sheridan, former Director of the MOA, thought the panellists were constructing through writing for MO (Sheridan, 1993). Hinton (2010, 2013) has demonstrated how the biographies of panellists can be studied and used to illuminate historical processes that are molecular and shaped by the choices of these historical agents. Hinton has also suggested, however, that volunteer writing for MO provides researchers with access to the cultural worlds inhabited by the panellists, the worlds of discourse – populated by newspapers, advice manuals, films and so on – from which people in particular places at particular times construct their selfhoods.

This brings us to an alternative way of approaching MO sources, which might be called ‘reading horizontally’ for shared cultural resources. We have described this approach at length elsewhere (Clarke et al., 2018). It is based on the claims of cognitive science that behaviour is shaped by understanding, which in turn is shaped by cultural models; the claims of social theoretical writing on discourse that social reality is constructed from practices, which in turn are shaped by forms of consciousness; and the claims of interpretive social science that actions follow from holistic beliefs, discourses and traditions. MO sources are read for the cultural resources panellists use to construct their understandings, including categories, storylines, folk theories and subject positions (Clarke et al., 2018). Examples from the present article include the norm of ‘responsibility to vote’ and the understanding of political campaigns as being made up of ‘claim and counter-claim’. This approach focuses especially on the cultural resources panellists share with each other and, plausibly, other citizens in their families, friendship networks, workplaces and audiences for cultural products. As such, in contrast to the vertical approach, reading horizontally privileges generalisation over difference, complexity and the context of individual panellists’ total engagements with MO. These strengths and limitations are shared with similar approaches taken by other researchers, for whom MO materials have disclosed ‘public and shared understandings’ (Salter, 2010), ‘shared cultural repertoires’ (Nettleton and Uprichard, 2011) and the ‘proverbs, truisms, and everyday episteme’ from which citizens construct and express opinions (Gazeley and Langhamer, 2013).

In the present article, we follow Clarke et al. (2017) in reading MO sources not only for shared cultural resources but also for the logical connections between phenomena they sometimes disclose. In the sections that follow, logical connections disclosed include those between demands that citizens express an opinion/vote, uncertain but dutiful citizens (section ‘Reluctant, Uncertain, But Dutiful Citizens’), unhelpful campaigns (section ‘Unhelpful Campaigns’) and opinion/vote formation by recourse to feelings and elite cues (section ‘Recourse to Feelings and Elite Cues’).

Reluctant, Uncertain, but Dutiful Citizens

When asked by MO if they would vote in the 2016 referendum, most panellists responded positively. They would vote, but the reasons given for why they would vote and the views expressed about the ‘privilege’ or ‘debt’ or ‘responsibility’ of voting are interesting and worth discussing. Many said they would vote not because they held strong views about Britain’s membership of the EU, but because they held strong views about voting. A sales ledger controller in her 40s from Northern Ireland (C5692) wrote: ‘I will definitely be voting [. . .] I personally feel very strongly that we must all vote’. A teacher in her 60s from Scotland (W729) would vote even though she was ‘sick’ of voting after the last few years of British politics: ‘I’m sick of elections for this, that, and the next thing, but I will vote because I think it’s important to vote’. Why was it important to vote? For many panellists, voting was viewed as a right won by their predecessors. ‘I vote in every election because people fought and died for this vote’, wrote a self-employed woman in her 40s from Northeast England (R5429). ‘I remember what the suffragettes did to get the vote for women’, wrote a shop assistant in her 80s from Yorkshire and Humber (J1890).

Elderly women especially would reference British history and the suffragettes in their responses. Other panellists would give similar justifications for voting, but were more likely to mention the lack of voting rights elsewhere in the world at the present time. A self-employed professional in his 40s from Southeast England (D4736) would vote because ‘We have the extravagance of freedom that many others don’t have in this world; a freedom that people are prepared and forced to die for, and we must not squander our privilege’. A pharmacist in her 50s from the West Midlands (V3773) would vote because ‘ordinary citizens of many countries still don’t have a vote’. A retired nurse in his 60s from the East Midlands (C4988) would vote because ‘I feel that living in a democracy which allows us to cast a vote is a privilege and one that many people around the world do not get to do’.

This last panellist went on to write: ‘I feel the referendum was called not because there was a pressing need for it’. He felt it had been called by Prime Minister David Cameron for internal party reasons and was not alone in feeling this. Many panellists admitted to feeling strongly about voting, but not about EU membership and related issues. For some like this administrator in his 40s from the West Midlands (B3227), the issues just did not seem that important: ‘The key issues seem to be the economy (including jobs and prices), crime and terrorism, migration and border control, and independence/lack of control over our own laws. I can’t say I feel worked up about any of them’. For others, the issues may have been important, but the question asked by the referendum would do little to address them. On this, a charity worker in his 30s from Northeast England (J5734) wrote: ‘I don’t see a great deal at stake. Ultimately, I’m deciding whether a remote and unrepresentative political class in Brussels get to play dress-up with other people’s lives, or whether a remote and unrepresentative political class in London get to’. Similarly, a self-employed professional in his 40s from Southeast England (D4736) wrote:

Key issues for the UK, I believe, are immigration, sovereignty, and how our money is spent [. . .] Leaving the EU is not a solution to any problem in itself, nor is staying [. . .] I can honestly say that I’m not too bothered which way the referendum goes.

If some panellists struggled to feel strongly about the question being asked by the referendum (but would vote), then others struggled to feel confident in their response to the question (but would vote). In the terms of Leruth and Stoker (2020), many panellists found the task environment of the referendum to be difficult. An apparently easy binary choice was presented, but regarding a seemingly difficult technical issue. One such panellist was a machine operator in her 20s from Northwest England (B5702). She would ‘definitely vote’, but was not sure she was ‘informed enough’ and was ‘not 100% sold on either side’ of the argument. Another was a finance manager in his 30s from Southwest England (T5672), who was ‘very much on the fence’ and ‘could genuinely see advantages both ways’. In the end, he voted ‘on balance’ and ‘after much hand-wringing’. He left his decision late, as did many other panellists. A retired library assistant in her 70s from the East of England (H2639) wrote that she intended to vote but had ‘not decided’ how. There were ‘so many issues’ and ‘so many opinions on these issues’. She wrote: ‘I shall definitely vote, but until I actually have the voting paper in my hand, I still cannot honestly decide which way I shall vote!’. Similarly, a retired nurse in her 80s from the East Midlands (M2061) wrote of the ‘pros and cons’ she had listened to during the first days of the campaign, before writing: ‘Up to now, I am in a complete fog as to which way to vote’. Later in her diary, she wrote: ‘I did not make up my mind until I was in the polling station’.

The panellists quoted so far in this section were voters in the referendum, but they were reluctant, uncertain voters. Some did not see the referendum question as important or consequential. Others did not see an obviously correct answer to the question. We complete this section by noting that some panellists wore this position of the reluctant, uncertain voter with apparent ease, but others appeared to feel its weight. A shop assistant in her 80s from Yorkshire and Humber (J1890) wrote:

It is very difficult to have enough knowledge and information about the EU. I keep changing my mind. I hear one thing which sways my decision, and then I hear another point of view and it changes my mind again. It really is a very big responsibility.

Just before polling day, she continued: ‘The 23^rd June is decision day, but I still don’t know what to vote [. . .] It is a big decision and what if I make the wrong decision?’. The responsibility of voting weighed heavily on this panellist. Other panellists went further in arguing that citizens were being made responsible for something that really should have been the responsibility of politicians. Consider the following three quotations, first from a civil servant in his 50s from Southwest England (E5014), then from a self-employed woman in her 50s from Northeast England (R5429) and then from a community health worker in her 40s from the East Midlands (T4715):

I think I will vote [. . .] but to be honest I don’t think I really know all the implications of either option [. . .] I am really annoyed that the Conservatives have left such an important decision to the electorate [. . .] I see the Government as abdicating its responsibility to govern.

I will vote [. . .] I firmly believe this issue should not be decided by the public [. . .] As a democratic country with an elected government and officials, I believe the issue should have been addressed by those people and consider it a ‘cop out’ by asking the public to sort it out for them.

I will vote, even though I don’t want to have to be involved in such a monumental decision. The politics of this issue are very intricate and I’m finding it difficult to get a clear understanding of it all. Politicians are paid and elected to make these decisions. I know they don’t always get it right, but it should be their job to sort these things out, not ordinary people.

These panellists would vote in the referendum, but would do so reluctantly. They assessed themselves to lack sufficient knowledge and understanding for such a decision. Their folk theory of democracy is reminiscent of what Hibbing and Theiss-Morse (2002) called ‘stealth democracy’. They saw their own democratic role as limited. They expected politicians to govern and take responsibility for issues like EU membership on their behalf.

We are now in a position to summarise our first claim. Many voters in the EU Referendum were reluctant and uncertain. They did not begin 2016 as Leavers and Remainers with strong opinions on British membership of the EU. They did not feel well qualified to make a decision. They felt obliged to develop opinions when asked to do so, not least because of the strong opinions they held about voting, which they saw as a debt to be paid to the suffragettes, or a privilege not to be squandered. These citizens ended up voting, but often after hesitantly developing and changing their position over time, and without feeling confident in the final position they had reached. Many resented being asked to navigate such a difficult task environment and take responsibility for such a difficult decision.

Due to the constitution of the MO panel, this material does not allow us to say how many voters nationally were reluctant and uncertain, as described in the paragraph above, but we argue this position was not unusual and so should be taken seriously. The quotations used above are just a few examples of many such quotations in the MO material. They are not taken from a narrow section of society, but from men and women of all ages and all regions of the UK; people working as teachers, shop assistants, pharmacists, machine operators, finance managers and so on; and citizens who, when push came to shove, voted Leave in some cases and Remain in others. The cultural resources these panellists drew on when writing their responses to MO – including the category of the vote as ‘debt’, ‘privilege’ and ‘responsibility’, or the storyline that ‘both sides’ had ‘pros and cons’ to be weighed ‘on balance’ – were shared by a wide range of panellists. They were presumably circulating widely in British society at the time. Therefore, it is plausible that such cultural resources were used by many citizens, far beyond the MO panel, to construct their opinions and votes.

Unhelpful Campaigns

In some ways, the panellists quoted in the previous section sound rather trusting of politics. They trust in democracy, in so far as they see voting as important. They trust politicians and government, in so far as they see issues like membership of the EU as best decided by citizens’ representatives, and not citizens themselves. However, there is one part of politics in which most MO panellists clearly did lack trust, or came to distrust over time: the referendum campaign – and this was important because how else were uncertain, ill-informed, hand-wringing citizens meant to form opinions and decide how to vote if not by engaging with the campaign?

Many panellists looked to the campaign for help, at least initially. They read leaflets and newspapers, watched television and listened to the radio. Fairly quickly, however, they came to realise that the campaign would not provide them with what they were seeking. They wanted facts about what would happen in the event of Britain either leaving or remaining in the EU, but they came to realise that such facts were not available. Only speculation was available. A self-employed professional in his 40s from Southeast England (D4736) bemoaned the lack of ‘factual evidence’ and the ‘speculation’. For him, this speculation was no better than ‘misinformation’. For other panellists, including a carer in his 30s from Northeast England (N5744), speculation was usually ‘mere speculation’ (emphasis added). A community health worker in her 40s from the East Midlands (T4715) was disappointed that ‘neither side can categorically and truthfully say what will happen’. The problem was put clearly by a student in his 20s from Northwest England (S5780): ‘It is impossible to have facts about the future. There can only be speculation’.

A storyline repeated time and again across the MO panel was that ‘no one knows’ what will happen as a result of Britain remaining in or leaving the EU. After following the campaign, a retired film writer in his 70s from Scotland (H1541) concluded that ‘no one really knows anything’. This phrase was repeated 3 times at different points in just this one campaign diary. Variations of the phrase were repeated by other panellists. ‘There’s simply no way of knowing what will happen if the UK votes for Brexit’, wrote a civil servant in his 50s from Northeast England (M3190). Also in her 50s, a retired teacher (M3412) wrote: ‘no one has any real knowledge or true idea of just what the outcome would be if we left. How can anyone know?’. Also now retired, a former journalist in her 70s from Northeast England (W633) wrote: ‘No one will know just how, for example, exports, currency movements, and trade treaties will be affected by leaving unless and until we actually do so [. . .] Brexit is a pig in a poke, but so too is Remain’.

Another common storyline, shared by many panellists, was that facts and knowledge in the campaign had necessarily been replaced by speculation, but also claims, counter-claims and accusations (usually of scare-mongering) – few of which could be verified and trusted. Consider this from the referendum diary of a student in his 20s from Northwest England (S5780):

The problem is that I just don’t know what’s true. [Andrew] Neil says that we couldn’t nationalise the railways if we stayed in the EU. [Hillary] Benn says we could. Neil suggests TTIP might threaten the NHS. Benn says that it won’t.

And consider this from the diary of a retired film writer in his 70s from Scotland (H1541):

The reaction by the Leave campaign to the Treasury’s 200-page report was incandescent and much use was made of allegations that ‘the fear factor’ and incomprehensibility were replacing reality and facts. The attacks were led by Justice Secretary Michael Gove, whose contributions were just as instantly condemned by Labour’s First Secretary Angela Eagle as ‘complete utopian rubbish’. [Dominic Grieve] appeared on The Today Programme on Radio 4 shortly after Michael Gove to denounce his claims about the European Court of Justice undermining the UK’s security as ‘unfounded and untenable’ [. . .] Just a few weeks ago, a former head of MI6, Sir Richard Dearlove, argued that ‘Brexit would bring two potentially important security gains [. . .]’. This weekend, two other former security chiefs – Sir John Sawers of MI6 and Lord Evans of MI5 – have pitched their tent on the other side of the barricade.

These diarists were describing what another panellist, a retired journalist in her 70s from Northeast England (W633), called ‘Newton’s law of politics: for every claim, there is an equal and opposite claim’. Most of the MO panellists noted this characteristic of the campaign. There were ‘claims and counter-claims’ that became ‘increasingly hard to decipher and believe’ (H1541), ‘endless claims and counter-claims’ that ‘make it impossible to detect the truth from the lies’ (S5767) and ‘claims and counter-claims becoming ever more hyperbolic’ (M3190). Many diaries were structured to reflect this: ‘On the one hand [. . .] But then on the other hand [. . .] Who is to be believed?’ (M5770). ‘One local MP (Labour) says [. . .] But another local Labour MP has said [. . .]’ (V3773). Panellists generally found this ‘to and fro’ (R3032) character of the campaign frustrating. ‘Each time one politician [. . .] makes a point’, wrote a retired teacher in her 50s (M3412), ‘there is someone from the opposition who is more than ready to ridicule the point’. Similarly, a retired civil servant in his 70s from Wales (R3032) wrote: ‘As soon as one side of the argument produces, for example, a list of prominent business people who support their side, the other side produces their list’. ‘Who do we believe?’, asked our retired journalist in her 70s from Scotland (W633), when ‘figures are quoted by both campaigns [. . .], each contradicting the other’.

It would help if there had been what many panellists described as a source or mediator that was independent, reliable and impartial – but apparently none could be found. A carer in his 30s from Northeast England (N5744) wrote: ‘It would be nice to find reliable and unbiased sources of data about the subject so that I could make an educated guess’. Also from Northeast England, a retired civil servant in his 50s (M3190) complained ‘that there’s no completely disinterested and impartial person, institution, or corpus of knowledge out there to whom appeals for advice and accurate information on the merits and drawbacks of EU membership can be directed’. For our retired film writer in his 70s from Scotland (H1541), not even the BBC was playing this role. It just reported ‘bipolar partisan claims without question’, such that ‘the listener is left with views from completely opposite viewpoints, with little exposure to mediation or moderating enquiry’.

Where did all this claim and counter-claim leave our already reluctant and uncertain panellists? They were left with ‘pros and cons for each opinion’, but ‘no conclusive arguments from either side’ (E5014). They were given ‘arguments for and arguments against’, but ‘no clear better argument’ (S5816). Finding themselves with ‘no right or wrong answer’ (N5744) and no ‘smoking gun’ (E5014), how did panellists respond? Many remained uncertain in their opinions. ‘I just don’t know what to think’, wrote a student in his 20s from Northwest England (S5780) towards the end of his diary. Another student in his 20s, this time from Southeast England (M5770), recounted the competing claims he had heard, wrote ‘it is hard to know what to believe’, and concluded just before polling day: ‘It is for this reason that I am unsure which side I want to vote for’. Some panellists despaired at being left in this position and the quality of the campaign. Long before polling day, a teacher in her 60s from Scotland wrote (W729), ‘I’ve lost the will to live. I am sending in this Directive now, not later. Sick, sick of politicians and their claims and counter-claims’. Of the campaign, with its apparently endless diet of speculations and accusations, a charity worker in his 30s from Northeast England (J5734) wrote:

I have grown old in its service. And now it seems that all my children will ever know is the sound of grown men yelling ‘Hitler’ at each other and debating just how many hypothetical beans hypothetical cows will fetch in an abstractly imagined future [. . .] Could we all just put rocks in our pockets and walk into the sea?

This panellist arrived at polling day in despair. Other panellists responded to the unhelpful campaign by disengaging. After all the ‘lies and misinformation’, a self-employed professional in his 40s from Southeast England (D4736) described himself as ‘baffled and disengaged’. Also in his 40s, an administrator from the West Midlands (B3227) reported how ‘the constant reiteration of each side’s accusations and counter-accusations (many of them hypothetical) has had a numbing effect on me’. Signing off his referendum diary, a civil servant in his 50s from Southwest England (E5014) wrote, ‘I’ve finally concluded that it doesn’t matter which way the vote goes’.

And yet we know that most panellists did end up voting, just like most citizens beyond the MO panel – many of whom, it is reasonable to assume, were also familiar with categories like ‘facts’, ‘mere speculation’ and ‘claims/counter-claims’, and storylines including the one about no one knowing what will happen, the one about it being impossible to know what to believe and the one about there being ‘pros and cons’ for each side and no clear best option. These were the prominent cultural resources offered to citizens by a context of low political trust during the 2016 campaign. So the question remains: if not by calculation – by slowly establishing the facts and assessing the costs and benefits – how did voters in the referendum eventually form their opinions and determine their vote?

Recourse to Feelings and Elite Cues

We have seen that many panellists felt a responsibility to vote in the 2016 referendum, but were uncertain of their opinion on Brexit and found the campaign unhelpful when seeking to perform this difficult task. We now turn to how some panellists finally managed to develop a position, allowing them to act – though still with uncertainty and hesitancy in many cases. If panellists could not know the consequences of voting Leave or Remain (‘no one knows’), they gradually realised they could know other things on which to base their vote – not as a preferred mode of decision-making, but as a last resort. Specifically, they could know their own feelings, beliefs, principles and visions for the future, and they could take a view on who they wanted to be aligned with – which politicians and other prominent figures in the campaign – and, crucially, who they were against.

Let us begin with feelings, a term we use loosely to capture the range of ways panellists made their decisions on the basis of not carefully weighed information about consequences, but a rather vague sense of what feels right, what is right to believe and what broad principles are right to uphold (cf. Moss et al., 2020). Sometimes, panellists used the term ‘feeling’ itself. For example, an engineering works manager in his 50s from Southwest England (S5915) wrote this when justifying his Leave vote: ‘My basic wish was to regain what I felt was a sense of destiny, rather than being forced to accept the resolve of unelected people in Europe’. Similarly, an author in her 30s from Scotland (J4793) wrote this when justifying her Remain vote:

I’ll admit to not knowing so much about all that goes on within the EU and all that it involves for a member state, but [. . .] it feels better to be part of something, contributing to something bigger, than peering in from the outside [. . .] I voted for unity and shared resources.

These panellists were falling back on rather abstract notions like destiny and unity. Sometimes, they discussed this in terms of ‘beliefs’. Here are two Remain voters on why they voted that way: ‘I voted Remain [. . .] My belief that nations ultimately work better when they are united together led to my vote’ (B5178); ‘I voted Remain. It wasn’t influenced by a particular issue but it was an inherent belief that we should be part of something bigger’ (H5210). Alternatively, they discussed it in terms of ‘principles’ and what is generally right and wrong. Here are two Leave voters on why they voted that way: ‘I recall being distinctly unimpressed by the campaigning of either side. I could see pros and cons of each outcome [. . .] In the end, I think my main influence was principle: simply the principle of self-rule’ (T5672); ‘The most important issue that influenced my vote [. . .] was that I felt it’s wrong to pay someone in Brussels to tell me what I can and can’t do’ (R3546).

Panellists were looking around for something to guide their opinion and vote. They were struggling to find what they wanted: reliable facts about what would happen in the event of Brexit or Remain. They were, however, finding something else. These could be feelings, beliefs or principles. They could be something rather modest like the rule of thumb used by this self-employed professional in his 40s from Southeast England (D4736): ‘There is so much uncertainty about the outcome of leaving that I would urge caution and stick with it. Better the Devil you know’. Or they could be something rather grand like the ‘vision for Britain’ used by this retired nursery teacher in her 60s from the West Midlands (M3408):

Nobody deals in facts about how this country may look in five or ten years’ time, because nobody can predict the future in regard to any of the major issues at stake [. . .] I haven’t actually learnt anything concrete about the better choice. I have, however, learnt a lot about what my vision for Britain in the future would be and which side of the argument is most in tune with my hopes.

This panellist ended up voting Leave, but she did not begin 2016 as a Leaver and she did not vote Leave on the basis of ‘concrete facts’. Like the other panellists quoted in this section, she felt an obligation to form an opinion and vote, and, finding the political information provided by the campaign unhelpful, settled on a different way of determining her choice. This alternative way for many panellists involved clarifying their general beliefs and principles – researchers may refer to these as ‘values’, but the MO panellists did not use that word – before feeling the fit between those criteria and their sense of Leave and Remain as positions, worldviews and tribes.

This brings us to elite cues because one of the ways this fit could be established – could be sensed as being ‘in tune’ – was by using the personalities involved in the campaign as a simple and quick proxy for the two positions on offer. In a context of low political trust, there were few if any panellists who followed the endorsements of politicians and voted a particular way. Instead, there were examples of panellists who followed the endorsements of non-politicians, such as this pharmacist in her 50s from the West Midlands (V3773):

I haven’t read all the leaflets or listened to all the comments [. . .] because I don’t actually trust most of the leading campaigners or believe that any of their ‘facts’ are true, whichever side they’re on. So I have to look at what other people have said, and when I look at those who are in favour of remaining, they tend to be people I trust more than those who want to leave.

The ‘other people’ referred to by this panellist included science writer Ben Goldacre and journalist Martin Lewis. The panellist did not write about their ‘facts’ or arguments for remaining in the EU – their message. She wrote about how she trusted them as messengers and voted Remain on that basis.

This panellist took what might be called ‘positive elite cues’ from non-politicians she trusted, but she also took ‘negative elite cues’ from campaigners, including politicians, she felt repelled by – an approach that was prevalent across much of the MO panel. She wrote: ‘Most of the better known fans of Leave seem to be rich or self-interested. Whatever their arguments (and I only have the gist of most of them), I just don’t think that they’re much like me’. Other panellists were also put off by those campaigning for Leave, who seemed unfamiliar, ‘shifty’ or worse – largely regardless of their ‘arguments’. ‘I was going to vote Brexit’, wrote a museum manager in his 60s from the East of England (A6056). ‘Then I noted that shifty characters like Farage, Gove, and Johnson were in favour of leaving, so I voted Remain’. Observing the Leave campaign, a retired film writer in his 70s from Scotland wrote: ‘I cannot imagine supporting anything they promoted’. This approach worked for panellists who voted Remain and also for panellists who voted Leave, including the civil servant in his 50s from Southwest England (E5014) who explained: ‘I found [David] Cameron and [George] Osborne the two most odious people alive, and the thought of upsetting their stupid little public school spat with their fellow Tories was very appealing’. The negative-cues approach only hit problems when distrustful panellists found themselves pushed in two directions by endorsements from both sides. ‘My first thought is that if David Cameron and his cronies are in favour of staying in the EU, then the general public would be wise to vote to Leave’, wrote a carer in his 30s from Northeast England (N5744). ‘Then again’, he continued, ‘Boris Johnson [. . .] is arguing that we should leave’. On the basis of negative elite cues, this panellist did not know which way to turn. They did, however, know one thing: ‘I guess what I’m saying is that my decision has been swayed on which way to vote by looking at which people are in favour of staying or leaving, and considering what I think of them’. As a reluctant, uncertain citizen, who nevertheless felt obliged to participate in the referendum, his vote had been swayed not by reliable political information, which the campaign had failed to provide, but by a series of judgements about individual politicians and other campaigners. Were they trustworthy? Were they like him? Did they share his values? These things could be known, in so far as they could be felt, even if the future consequences of Brexit or Remain could not be known.

Discussion

Our main aim in this article has been to understand more about how citizens form opinions and come to decisions as voters in a context of low political trust. To do this, we have used volunteer writing for MO. The diary format used by many panellists in response to MO directives on elections and referendums allows researchers to reconstruct what happens between the earlier demands made of panellists to think about an issue and the later act of voting. These diaries, we have argued, tell us about the understandings, contexts and practices of MO panellists, but also many other citizens beyond the MO panel – so long as the panel is sampled to cover a broad range of geographical, political and social positions, and the writing is analysed for cultural resources used by panellists from across these positions, indicating that such resources were circulating widely in society at the time in question.

By tracking the formation of public opinion, we found that a lot happens between the earlier demands made of citizens to think about an issue and the later act of voting. In discussion groups held just prior to the 2016 referendum, Andreouli and Nicholson (2018) found that many citizens were forming and reforming their positions by wrestling with dilemmas. Similarly, in MO sources, we found that many citizens developed opinions reluctantly, hesitantly and uncertainly. They lacked strong views about issues like the EU and Brexit, but held strong views about voting, which they saw as a debt, privilege and responsibility. Feeling this responsibility and finding themselves in a challenging task environment like the EU Referendum, many citizens look for information so they can view themselves, and be viewed, as rational choosers. However, in a context of low political trust, campaigns often do not function to provide such information. Citizens want facts but receive ‘mere speculation’. Claims come accompanied by counter-claims. ‘Pros and cons’ build up for each side. Nothing conclusive is provided by campaigns.

In such cases, many citizens form opinions and make decisions by falling back on rather vague feelings about their own beliefs, principles and visions, and the fit of these feelings with particular campaigns and especially campaigners. On both feelings and elite cues, we make an original contribution to the literature regarding their role and functioning in a context of low political trust. Taking feelings first, in the literature on Euroscepticism (e.g. Hobolt et al., 2011), it is often claimed that Euroscepticism has become increasingly based on political-cultural feelings instead of cost–benefit calculations because the EU has developed from a free market into a political union, which has brought questions of sovereignty and national identity to the fore. This may be so, but our study has found another explanation for the reliance on feelings by citizens taking a view on EU membership and many other issues in recent years. In a context of low political trust where facts are disputed, predictions of consequences are dismissed as mere speculation or fear-mongering and it is difficult to know who to believe, one thing a citizen can know is their own feelings – especially regarding their own beliefs, principles and visions, and especially regarding individual personalities associated with certain positions and campaigns.

Turning to elite cues, most of the literature on these particular heuristics assumes a context of high political trust and a positive relationship between elite endorsements and voting patterns. This was the context for King’s (1977) study of the 1975 Referendum on the Common Market. At a time of negligible ‘anti-establishment mood’, he found that voters ‘followed the leader’ and backed ‘Yes’ because that was largely what the two main parties – and especially the party leaders –recommended. By contrast, in the current period of low political trust, we found something quite different. We found evidence that elite cues drove decision-making in the 2016 referendum – something also found by other studies (Clarke et al., 2017b; Hobolt, 2016). Beyond that, we found that negative elite cues were particularly important in driving votes for both Leave and Remain. Voters used particular politicians as proxies not for positions they wished to support, but for positions they wished to vote against.

Reflecting on these findings and the literature reviewed at the top of this article, it seems that a context of political distrust favours fast-thinking heuristics over slow-thinking cost–benefit analysis. In the terms of Hooghe and Marks (2005), it favours ‘community’ and ‘cues’ over ‘calculation’. This holds if we take ‘community’ to mean feelings of national identity and belonging, as Hooghe and Marks do, but also feelings regarding beliefs, principles and visions more generally. It also holds if we take ‘cues’ to mean negative elite cues, as opposed to the positive cues of elite endorsements.

Finally, we consider the question of how the EU Referendum result and associated polling results might be interpreted. Since many opinions and votes were based on vague feelings and negative elite cues, and were arrived at reluctantly, hesitantly and uncertainly, claims regarding what people think about the EU or Brexit should be treated with caution. This includes the claims of populists about ‘the will of the people’, but also those of pollsters and political scientists about polarisation.

There are two main positions on polarisation in the literature. One is that long-term polarisation – social, economic and cultural – produced the conditions for the referendum and result (Ford and Goodwin, 2017; Goodhart, 2017; Norris and Inglehart, 2019). The second is that polarisation between Leave and Remain identities was actually produced by the referendum, campaign, outcome and aftermath (Curtice, 2018; Hobolt et al., 2020). Sobolewska and Ford (2020) combine these two positions. They argue that polarisation of ‘identity liberals’ and ‘identity conservatives’ was driven by long-term demographic and social changes, and exploitation of the issue of immigration by UKIP, followed by the referendum campaign and result, leading the UK to ‘Brexitland’ (a divided nation that became conscious of its divisions).

The immediate evidence from our study would seem to put our argument nearer the position that polarisation was produced by the EU Referendum. But we think it takes us beyond that position, with important implications for UK politics. We have evidence that people ended up voting for one particular side of the divide often because of strongly held negative views about the other. But we have little evidence that people held strong and positive views about their own side – or even that they saw it as ‘their own side’. A diagnosis of polarisation requires a divided people not only detached from ‘the other side’ but also attached to their own side. In the absence of strong in-group attachments, we suggest there is a lot of work being done by the generalised lack of trust in politics and that is key to understanding how UK politics will unfold rather than polarisation.

As noted, the 1975 Referendum took place in a context not of anger and resentment of the establishment, but of deference or acquiescence to leaders who were largely followed by voters. The 2016 Referendum took place in a very different environment. The strongest divide in British politics is between ‘us’ (the people) and ‘them’ (the elite), not the gap between Remainers and Leavers. The tactical game for leaders and parties is to focus on issues where us versus them can be mobilised in their favour. The Leave campaign did that more effectively in the referendum campaign through rhetoric such as ‘take back control’ and ‘people have had enough of experts’, and in framing of the post-referendum slowness of the UK Parliament to honour the result, culminating in Boris Johnson’s pledge to ‘get Brexit done’ in the 2019 election campaign – pitching the Conservatives against an out-of-touch parliamentary elite that was obstructing the will of the people. Britain’s politics is currently framed as much by who can mobilise anti-establishment and more broadly anti-politics sentiment, as by polarisation along cultural lines.