Gender , ownership and engagement during the European Union referendum : gendered frames and the reproduction of binaries

The European Union referendum was supposed to be a significant moment for political engagement and ownership in the UK. This article looks at how the two official European Union referendum campaigns (Vote Leave and Remain) framed discussions about the UK’s membership of the European Union, as well as the impact of the campaign on women’s political activation. Using data from a survey questionnaire conducted two weeks after the European Union referendum (in July 2016), we analyse women’s sense of political efficacy and engagement with European politics. We project those findings on a frame analysis, where we assess the footprint of each campaign in terms of issue coverage and the salience of gender as a campaign issue. Our findings shed light on the way in which issue framing and confidence affect the quality of political engagement among ‘weak publics’.


Introduction
The 2016 European Union (EU) referendum was a critical juncture in UK and European politics, and not just in terms of the relationship between the UK and the EU or the rise of Euroscepticism across Europe.Gender scholars have been calling for the inclusion of a gender dimension to the debate about EU membership (Haastrup et al, 2016;Guerrina et al, 2018;Rubery and Fagan, 2018), and since the referendum, civil society organisations have been advocating for gender impact assessments and a commitment to women's rights in the context of the ensuing complex negotiations (Fawcett Society, 2018).This article contributes to the debate about gender silences in the way in which the UK's withdrawal from the EU was discussed before, during and after the 2016 referendum.Specifically, it draws on a growing body of research that points to the shortcomings in the way in which scholars of EU studies have approached the subject and how this process ultimately reproduces hierarchies of power within the discipline and the wider political environment.The EU referendum and the associated campaign should have been an opportunity for open and detailed public debate about issues relating to EU politics and policies, as well as the future direction of travel for the UK.Moreover, the high-profile campaign might have constituted an opportunity to close the gender gap in women's political engagement with high politics, and European politics specifically (Galligan, 2015), as well as gendering the Brexit debate.Despite expectations ahead of the 2016 vote that women would likely vote remain at higher rates than men, in actual fact, there was no gender gap in voting patterns (Ashcroft, 2016).This finding is in line with a growing body of literature on the limited opportunities offered by referenda as a tool for gender-sensitive constitutional renewal (Bell and MacKay, 2013;Kenny and Verge, 2013).
Given the importance attributed to the wider political context as a key determinant of political participation (cf Lowndes, 2004), this article explores the role of the official EU referendum campaigns on women's political engagement.More specifically, we explore the role of the campaigns in defining the parameters of political debate and, in so doing, reifying the high-low binary as linked to political salience rather than efficacy and engagement.Mapping the footprint of the campaigns in terms of issue coverage allows us to study how gendered narratives influence and affect women's engagement with the debate about the UK's membership of the EU.Three interrelated questions are addressed: (1) 'Were the official EU referendum campaigns gendered in their scope, focus and reach?'; (2) 'Which policies and issues did the 2016 referendum campaigns focus on, and what does this tell us about the biases underpinning political debate?'; and (3) 'Did the campaigns' strategic focus affect women's political confidence, especially women's engagement with the key debates in the EU referendum campaign and their tendency and direction of vote?'.In this way, the article seeks to map the effect of the EU referendum on women's sense of political ownership, understood to be key to generating a higher level of engagement (Warleigh, 2000).Specifically, it looks at the way in which the official 2016 EU referendum campaigns positioned issues relating to equality, women's rights and social cohesion within the context of the UK's relationship with the EU.
Underpinning this research is the assumption that while there was no gender gap in electoral turnout and voting in 2016, the deeply gendered nature of the campaign reproduced political binaries.The process and highly emotive nature of the debate ended up concentrating almost solely on what are perceived to be high-salience issues at the expense of social policy and politics.Consequently, the analysis presented here has important implications in so far as it provides an assessment of the gender bias in the framing of political campaigns, and its alignment with patterns of women's political activation and engagement.Moreover, it provides additional evidence of the marginalisation of issues relating to equality and social cohesion at times of constitutional transition underpinned by referenda, observations already mapped in relation to Scotland and Catalonia (Bell and Mackay, 2013;Kenny and Verge, 2013;Kenny, 2014).

Gender, political engagement and Europe: knowledge deficits in the public and private sphere
To understand the way in which women in the UK engaged with the 2016 EU referendum, we need to unpack the complex relationship between political activation, efficacy and ownership as applied to the way that this demographic group generally engages with EU politics.It is widely assumed that there is a gender gap in public opinion about the EU, with women being more Eurosceptic than men (Liebert, 1999;Lubbers and Scheepers, 2010).Liebert (1999: 206) found gender disparities existing as early as 1994 across most EU members.In the case of the UK, women score consistently lower than men in the level of attachment to the idea of European integration (Liebert, 1999).These lower levels of political engagement are compounded by the perception that women have lower levels of knowledge about the EU, its institutions and its core policy areas (Eurobarometer, 2001(Eurobarometer, , 2013(Eurobarometer, , 2015;;Nelsen and Gruth, 2000).Other research has noted that this gender gap has remained stable across time (Fraile, 2014).This is an interesting, and indeed surprising, paradox as equality between men and women has consistently been promoted as a foundational narrative of the EU (MacRae, 2012).While it is beyond the scope of this article to explore the nature of the EU as a gender actor or regime, it is worth noting here as this might have provided a platform for detailed discussion about the UK's priorities in this area post-Brexit.
The historic data on women's engagement with EU politics and policies provide the backdrop to the research puzzle explored in this article, which focuses on scoping women's pathways to engagement in the 2016 referendum.Here, we want to think about how women's political efficacy is affected in the context of a highprofile single-issue campaign.This is not an uncommon preoccupation in relation to referendum campaigns; it has been mapped by Kenny (2014) and Verge et al (2015).Understanding the factors influencing women's public and private engagement with highly salient issues helps us understand how the Brexit campaign reproduced gender gaps in political ownership and confidence.It also provides important additional evidence to current debates about how referenda affect political activation more widely.Verge et al's (2015) analysis of gender, risk aversion and political change is particularly useful here.They find that in the context of 'high-stakes' political decisions, such as the independence votes in Catalonia and Scotland, political confidence and activation rely on the representation of women's interests during the campaign.Much like in the 2016 EU referendum, it is interesting to note the lack of a gender gap in the Scottish and Catalan referenda a few years earlier.This is something that requires further exploration, and may well relate to the nature of the binary question posed by referenda, women's pathways to political activation and the nature of the campaign, particularly when focused on high-salience/high-politics issues, for example, migration and security.
Political theorists, as well as social psychologists, have been working on defining or isolating variables that can help to explain levels of political engagement among groups and individuals.Theoretical explorations range from a focus on the individual -for example, rational choice theory -to the impact of policy outcomes and political structures on groups' participation.Issues of ownership and engagement at different levels of governance are enmeshed within this discussion (Barnes et al, 2004;Cicognani et al, 2009).These discussions build on many other studies that find long-term disparities in men's and women's political knowledge and confidence.Research also finds that such gaps lead to lower levels of efficacy and engagement in political discussions (Dolan, 2011).As women have been found to have consistently lower levels of attachment towards the EU, coupled with a sustained knowledge gap, it follows that they are also more likely to see the EU as not relevant to them as it is perceived to be distant/removed from everyday life.Accordingly, understanding how women engage with the information and ongoing debates about the future of Europe is an important story to be told about the 2016 EU referendum.This is a particularly important consideration in so far as it highlights the influence of knowledge, individual confidence and representation in shaping political attitudes in the context of a 'high-stakes' single-issue vote that requires voters/citizens to engage in detailed discussions about 'high politics' and technocratic issues (Verge et al, 2015).
Unpacking the complex relationship between the nature of the 2016 referendum, the UK's future relationship with the EU and the impact of Brexit on the lives of women in the UK hinges on individuals' confidence to engage with the issues.Starting from public opinion surveys reporting that women were consistently more likely to register as undecided about their voting intention, it is important to understand if, and how, the Brexit campaign sought to engage women as one segment of the electorate.To explain women's engagement with the EU referendum, in other words, we need to understand how women engaged with the information that they were provided with to decide which way to vote, as well identifying with whom they discussed issues relating to the vote and how this helped to inform their decision.
Studies into the nature and impact of the gender gap in voting behaviour/patterns frequently concentrate on voting preferences and political alignment (Inglehart and Norris, 2000;Nelsen and Gruth, 2000).Feminist scholars have also highlighted the complex relationship between campaigns, voting realignment and descriptive representation.Exposing when 'women's issues' enter the agenda, who discusses them and the impact on the 'high-level campaign' provides important insights into the work that gender is doing in shaping women's participation and engagement (Campbell and Lovenduski, 2005).There are two main approaches.First, feminist researchers have sought to expose bias and silences in the way in which campaigns engage with and represent women's interests.Examining the relationship between increasing women's political representation and the 'feminisation' of politics, this group of scholars seeks to highlight the permanence of binaries in the way in which politics is conceptualised, framed and practised (Lovenduski, 2005; for a sociological perspective, cf Beck, 2015).Second, others, preferring to treat gender as a variable, explore the role of attitudes and seek to account for the continued presence, decline or absence of a gender gap.
In the context of the UK's EU referendum, the official campaigns largely failed to give voice to and/or engage with women and 'women's interests' during the official campaign period.This was despite widespread media coverage that women were more likely to answer 'undecided' before the vote, and thus accounted for the largest proportion of swing votes (Katwala and Ballinger, 2016;Creagh, 2016) (a finding consistent with Fraile's [2014] cross-national study).This failure nevertheless reveals a deeply gendered campaign.Commentators' failure to predict women's voting intentions highlights how most public opinion studies have sought to understand this gender gap; it betrays an ontological bias about the researchers' understanding and treatment of 'gender' as a variable, rather than a structure of power.It thus only seeks to explain women's attitudes, rather than unpack the complex set of hierarchies that inform, and influence, women's political ownership and engagement.
The failure of both campaigns to address issues relating to social policy, and the under-representation of women's voices in media coverage, are two of the most significant silences of the research industry that has emerged around Brexit.Issues relating to women's rights, gender equality and employment policies were largely ignored by political leaders, campaign activists and researchers alike (Haastrup et al, 2016;Hozic and True, 2017).From a substantive representation perspective, this is all the more peculiar considering the impact of EU policies on the development of a national equality framework in the area of employment rights (Guerrina and Masselot, 2018).It would have been expected that at least the Remain campaign would have sought to use this platform to reach out to women voters during the campaign.As our analysis of the campaign material highlights, this was just not the case.
A 2016 report by the University of Loughborough Centre for Research in Communication and Culture found that in terms of gender balance, media coverage was substantially skewed in favour of men, both politicians and experts (CRC, 2016).In particular, the report tracked the prominence of women on TV and the printed press during the official referendum campaign.What they found was a rather stark lack of balance.In Brexit coverage, women featured only 17.5% of the time: only 25% of the TV and 15% of the press coverage featured women.This lack of visibility of women is also reflected in the lack of representation of social policy and genderrelated issues in the official campaigns, as our analysis of the campaign material reveals later.Shocking as these figures might be, they should not come as a surprise.The 2013 Scottish referendum also saw the marginalisation of women's voices and issues during the campaign (Kenny, 2014).
The effective exclusion of women's voices from the Brexit debate significantly curtailed opportunities for women's political engagement, limiting the impact of this 'citizenship moment' for women (Guerrina, 2016;Haastrup et al, 2016;Kenny, 2014).Hozic and True's (2017: 271-2) analysis of the lack of representation of and for women is altogether damning: With respect to the causes of Brexit, we emphasise that the limited gendered representation in the public referendum debate, which foregrounded largely white, male elites and technocrats effectively evacuating diverse voices including women's of all classes and races, led to an outcome unreflective of the range of views and experiences from within and across social groups.
The marginalisation of women's voices, they argue further, had the impact of othering women and limiting their presence in what they call a watershed moment for British and European politics.This process has permeated into wider Brexit research, with gender-sensitive analyses remaining largely absent from the scholarly discussions seeking to understand how Britain voted in 2016 (Gidron and Hall, 2018).
The focus of both the Leave and Remain campaigns on what were deemed to be 'high-salience' issues helped to crystallise the high politics-low politics binary.The limited number of voices contributing to this discussion compounded this further as the campaigns symbolically pushed women to the margins of the public sphere, thus implicitly recognising them as 'weak publics'.In this respect, the EU referendum campaign impacted the way we think and talk about citizenship.The focus on 'strong publics', that is, men, ultimately limits the reach of this moment in terms of democratic engagement with women, as well as the division between the public and the private spheres of life.This is a topic that has been discussed widely in feminist writings over the last three decades (see, eg, Landes, 1998;Celis et al, 2013), and despite widereaching changes in the legal status of women, it remains a fundamental concept in contemporary feminist analyses of social, political and economic structures.In this respect, the public-private dichotomy is an inherent feature of political relations, and as citizenship emerges from this context, it inevitably embodies this division (Lister, 1997: 66-72;Hantrais, 2000: 136;Vogel-Polsky, 2000: 62, 75;Mazey, 2001: 7).This broader discussion goes to the very heart of our article.As we explore the focus of the official campaigns in defining the boundaries of the political debate and post-referendum agenda, we also highlight the impact that this particular exercise in participation has on women's political activation and engagement.
Indeed, the fact that before the referendum, women were consistently more likely to say that they were undecided (Curtis, 2016) raises important questions about women's political activation during and after the referendum.This finding, in and of itself, is not particularly surprising and supports most of the work by gender scholars on women's political participation, whereby lower levels of confidence and knowledge in politics result in women voters' higher likelihood of answering 'don't know' in relation to voting intentions (Fraile, 2014;Galligan, 2015).There is a growing body of work highlighting the marginalisation of gender equality as a reference point in campaigns, particularly in the context of referenda.Kenny (2014), for instance, found that the issue of equality was largely absent from the 2014 Scottish independence referendum.This was something of a surprise given the centrality of the issue in public discourse around devolution.In addition to this literature, we disentangle the complex relationship between the nature of the 2016 referendum campaign, women's political ownership and the reification of the high-low binary at the heart of contemporary European politics.

Data and methods
Given the high visibility of the 2016 referendum campaigns, we were interested in systematically collecting evidence on political attitudes and behaviour in order to draw conclusions about connections between gender and responses to the EU referendum in terms of voting patterns, engagement with the question at stake and the salience of issues in voters' minds.At a second step, we examined the extent to which attitudes may have been embedded within an idiosyncratic political environment that was polarised or demarcated by the official referendum campaigns.To these ends, we operationalised our research questions in two ways: first, through a post-EU referendum public opinion survey; and, second, through a frame analysis of the referendum campaign material as emerged during the campaigns.
To assess political behaviour, we commissioned a survey with YouGov, just two weeks after the EU referendum vote (6-7 July 2016).This was intentionally designed to capture the emotional economy of the outcome of the referendum.We used a nationally representative sample of all UK adults (18+) weighted for those eligible to vote in the referendum.We had a total sample of 1,707 participants.The survey was administered online through YouGov's panels, with any and all shortcomings that this method may have (Fricker, 2008: 203-4).Beyond the question on voting preference in the referendum, we included questions on: the emotional state of play about the future of the country (Exadaktylos et al, 2017); the perceptions (negative or positive) about the campaigns, political parties and the government; issue salience in voting preference; sources of information; engagement with campaign activities; engagement with other people on the EU referendum; and regret of vote.We crossreferenced our survey results with the actual results of the referendum, finding that it was fought on very tight margins (Ashcroft, 2016) and across specific cleavages.
To test our assumption regarding the framing of the campaigns in terms of issue salience and the prominence of particular topics, we ran a thorough frame analysis of the campaign material published following the referendum's announcement (20 February 2016).We collected material published online by the two official referendum campaigns, Vote Leave and Stronger In.This would capture the official and formal lines of political communication without engaging with grass-roots activism or specific agendas.Material included information leaflets and press releases, campaign activity material, and publicised statements from the leaderships of the campaigns.We began collecting material from the day of the announcement and stopped collection on the day before the referendum.This resulted in a total of 1,092 relevant documents over the 18 weeks of campaigning.We coded at the document level for references to gender-framed policies versus traditionally gender-neutral policies.Gender-framed issues were identified as 'gender-equality policies' or those having a direct impact on women's access to the labour market and the public/political sphere.We coded for direct references to 'equal pay', 'maternity rights', 'work-life balance', 'women's rights' and 'women on boards'.We also coded for implicit gender frames, that is, issues identified in the literature as most likely to increase women's engagement with politics, traditionally associated with social policy and having a direct relationship with women's role in the private sphere.These included references to employment, health care, education, family finance/budgets, family rights and human rights.At this stage, the documents were already skewed towards the Leave campaign, with the Remain campaign producing only a quarter of the material that the Leave campaign produced (226 versus 866).This is itself is an interesting finding regarding the zealousness of the two camps.We will return to the aggregation of these results later.We had a team of four coders, with whom we ran a pilot study of 100 items, we then compared the scores and ran a reliability test with Krippendorff's α at 0.92.

Women's response to the EU referendum
The aim of the survey was to map the relationship between voting intentions, engagement and issue salience.We therefore focused on respondents' attitudes towards particular policy issues in deciding vote choice in the EU referendum.What results from this analysis is the observation that the main battlefield was between the economy (48%) and immigration/asylum (46%).As will be shown in the following section, this maps well on to the main focus of the campaigns and the increasing importance of these two areas in relation to voting intentions.Additional issues, such as Britain's position in the world (33%), Europe (25%) and security and defence (20%), were also identified as salient.On the other side of the spectrum, we found issues such as health (16%), terrorism (16%) and 'low politics' issues, such as the environment, education, crime, family life and childcare and women's issues, played a less important role.The particular framing of the campaigns -mainly around the economy at large and immigration -is reflected in the issues that framed voters' perception of the EU referendum.There is nothing particularly new here.It supports the findings of other studies that point to the salience of immigration and austerity, rather than 'Europe', on voters' choices (Hobolt, 2016).It should be added that this is not unique to this particular European referendum either; similar patterns are found in other, earlier, referendums as well (Garry, 2014;Verge et al, 2015).When we correlated gender as a variable with the responses in policy area importance, gender made no difference in the 'high politics' issues.Yet, as Table 1 shows, women were more likely to be influenced in their choice by 'low politics' issues that affect them individually, or their household, directly or indirectly.Notably, health, terrorism and women's issues are statistically significant for female respondents (at .01).This is an important finding as it supports earlier research about women's activation.
Health is especially interesting as a campaign issue as it was an important battleground for the campaigns.Additionally, health has strong links to the private/domestic sphere in so far as it relates to individual and family health, and it was also framed as having high salience by both campaigns.We will come back to this discussion later in this and the next section of the article.
Regarding the process of information gathering, we looked at which sources voters accessed to find out about political issues during the Brexit campaign.This is important as it speaks to issues of political competence, confidence and ownership.We were particularly interested in understanding how information shapes women's political engagement, recognising the long-standing knowledge deficit that women record in relation to the EU and European politics.The primary sources of information were television (67%), online (37%) and print newspapers (36%).Social media and experts were an important source of information (both at 30%), alongside friends (27%) and close family (24%).However, were there any sex differences in the way in which men and women engaged with any of those sources of information?Again, here we find a gender gap.As Table 2 demonstrates, women were more likely to receive their information via close family and friends, whereas men gained theirs through experts and analysts (in a statistically meaningful way, at .01).In other words, women were more likely to identify the private sphere/family as an appropriate source of information, whereas men were more likely to engage with official and public sources.This private-public divide is further accentuated when we zoom in on the people with whom our respondents discussed the EU referendum prior to voting.Perhaps unsurprisingly, the majority of people discussed the EU referendum with partners and spouses (53%) and close friends (52%), but less so with colleagues (35%), other family (34%), parents (32%) and children (26%).Nonetheless, as shown in Table 3, women were more likely to discuss the EU referendum in the private sphere, with their children, parents and other family members, compared with men.Men were more likely to discuss the issue with colleagues, either at work or outside work.This finding speaks to political efficacy and knowledge.It is also important as it compounds the marginality of women's voices in the media.
A critical question remains as to whether the sources of information for women and the people with whom they discussed the EU referendum affected their voting behaviour in the EU referendum.Table 4 paints a completely different picture.While particular sources of information had an effect on the way in which people voted in the EU referendum, there is no statistical significance when gender is introduced into the regression model.The same results appear when we introduce gender in the regression regarding the people with whom they discussed the referendum.Whereas the debates within our surroundings had a significant impact on the decision for either leave or remain, reinforcing the idea of echo chambers, gender had no effect.These are important observations that illustrate the absence of gender as a frame in both  'sources of information' and discussions taking place around the EU referendum.This analysis thereby shows that gender is significant in terms of ownership, particularly in relation to the respondents' confidence in their own knowledge of politics, even as it does not necessary affect participation and voting choices.Finally, whereas the knowledge gap on EU issues is an important observation in the Eurobarometer data, we understand that the EU referendum was not fought on the basis of knowledge of the EU.This is an important observation given the rank of 'Europe' as an issue in the earlier question on salience.It is also important to note (see Table 5) that the interaction of gender with EU knowledge had no statistical significance.This reinforces the argument that the campaign framings were not trying to inform citizens vis-a-vis their choice, but rather to capture voter sentiments on other issues.
Our new evidence confirms the original assumption of an absent gender gap in voting patterns in the EU referendum.However, it provides some useful insights into why this might be the case, namely, that the family has an important role to play in framing discussion and women's pathways for political engagement and activation.More qualitative research is required, however, to understand different patterns of influence on women's decisions in high-stakes elections.The campaigns constituted an opportunity to close the knowledge and engagement gap, particularly in the way that women engage politically with high politics and European politics.Yet, the survey indicates that the campaigns seemed to have had no effect in drawing in women more than men on these issues.In fact, the silence of gender as a framework for discussing the EU referendum overall, and the potential for this to swing the vote either for leave or remain, was coupled with the continued importance of the family as a framing and socialisation point for the political engagement of traditionally weaker publics, for example, youths and women (McDevitt and Chaffee, 2002).Again, more qualitative research is required to further unpack this trend and to understand in detail the opportunities and constraints for women's engagement and citizenship activation.

Unpacking gender frames in the EU referendum campaigns
To better understand the survey findings vis-a-vis women's engagement and activation, we analysed the online material produced by the two official campaigns: Vote Leave and Stronger In.How did each seek to mobilise the electorate?We looked at if, and how, women were 'mobilised/activated' by the campaigns, particularly given that women were the 'undecideds' before the vote.Based on the discussion of the survey earlier, our analysis focused on key policy areas, issues of salience and high politics to draw attention to key silences in public discussions about the nature of the EU, the impact it has had on national policies and the future relationship between the UK and the EU.Here, the focus of our analysis is to identify the scope and footprint of the campaigns in order to identify, first, the key frames that underpinned the public discourse and, second, whether these engendered higher levels of political ownership, thereby closing the knowledge gap.More specifically, understanding issue framing, salience and coverage helps us to provide some insights into the self-reported knowledge gap and lack of gender gap in voting outlined in the previous section.We collated all online campaign material produced by the official campaigns between January and 24 June 2016, thereby including promotional material produced ahead of the official six-week campaign.This is particularly important because most of the discussion of gender and/or women's rights issues in the EU campaigns coincided with International Women's Day (8 March 2016), which was outside the official campaign.We coded a total of 1,092 documents, of which 866 were produced by Vote Leave and 226 by Stronger In.Vote Leave had by far the largest footprint in the campaign, with 79.3% of the total campaign material.This finding reinforces other studies, for example, those looking at the scale and impact of social media engagement (Usherwood and Wright, 2017).
All documents were coded according to three broad categories: (1) genderequality policies or issues; (2) social policies; and (3) 'mainstream (high-salience)' policies and issues.These categories are based on a division of what are traditionally considered issues of high and low politics.Category one (gender-equality policies) included issues such as women's human rights, maternity rights, equality, women on boards and work-life balance.These are considered issues relating to the substantive representation of women's interests in politics and the economy.Category two (social policies) included issues such as employment, health care, family rights and education.Although these are not classed as women's rights issues, they tend to have a higher impact on women's political activation.Taken together, categories one and two are representative of policies often associated with 'low politics', that is, those that are often seen as less significant in the pursuit of the national interest.Finally, category three includes issues that have been identified as 'high salience', such as the economy, immigration, defence, security and trade.
Our analysis found that both campaigns concentrated on 'mainstream/high-salience' issues, a focus largely reflected in the overall media coverage during the referendum campaign (Moore and Ramsay, 2017).Taken together, the issues that received the greatest amount of attention were the economy, immigration and security.There are, however, some significant differences between the campaigns (see Table 6).Vote Leave did not attempt to engage in any meaningful way with issues relating to gender and women's rights, as evident in Table 1.Social policies were also very much relegated to second-order issues, at only 11.32% of the total coverage.Vote Leave's main attempt to engage with women occurred around International Women's Day, when the Women for Britain group (affiliated with Vote Leave) produced a video making an emotive plea to women to vote to leave the EU in order to safeguard the legacy of the Suffragettes (BBC, 2016).In terms of social policies, the main message of Vote Leave related to health care and employment.The National Health Service (NHS)-related campaign messages of Vote Leave were particularly memorable to more than one in four respondents in our survey (the largest share across all campaign themes), 26% for women.This is important in view of the results of our survey discussed earlier in this article: it is an example of an issue frame that can be seen as activating for women because it relates to the family as well as the individual.With health, and specifically the NHS, evidently becoming a high-salience issue during the campaign, additional research should investigate whether the issue was framed specifically to mobilise and engage women.Stronger In also focused on high-salience issues, especially the economy.Again, this is not a new finding and corroborates other studies (Usherwood and Wright, 2017).What is surprising, however, is the lack of detailed attention to social and equality policies/issues.Considering the potential impact of Brexit on women's employment rights (Fagan and Rubery, 2018;Guerrina and Masselot, 2018), we would have expected Stronger In to open up a space for detailed discussion of this policy area.Employment is the second most frequently mentioned issue after the economy, but the total footprint, and therefore potential impact on voters' activation, is minimal compared to the Vote Leave campaign.Lack of engagement by Stronger In with these issues, for example, exposing the impact of the repatriation of powers on women's labour market position, is perhaps one of the most pervasive silences of the whole campaign.
There is a clear distinction between the two campaigns.Whereas Vote Leave sought to highlight the potential for the UK, once it had repatriated powers from Brussels, to engage in a programme of deregulation and increased global competitiveness, Stronger In sought to focus on the economic threats to the British economy posed by the UK's marginalisation from the European market.Our analysis provides additional evidence for the argument put forward by Hozic and True (2017) that the language and focus of the campaigns and the ensuing Brexit narratives have helped to hide the impact of this process on the political economy of families in the UK.Here, we note a missed opportunity for Stronger In to engage with women voters.This could have been an activation point for women, but for this to take place, the issue of equality and social justice needed to be raised to the level of high politics.Such a discussion would have required a detailed engagement with the way in which institutional structures and dynamics have helped to establish a safety net for women employed in the labour market.In any event, these remained low-salience issues.Of course, had Stronger In offered such a detailed engagement, then this would have challenged the UK's ability to act independently in key areas of social policy (Guerrina and Murphy, 2016;Hantrais, 2018).This latter discussion goes beyond the scope of this article, but the broader point about how equality and social justice constituted low-salience issues is an illustration of how both campaigns failed to activate women as informed/knowledgeable citizens in the referendum.This takes us back to a point we made earlier in relation to 2016 being a missed opportunity for engagement and citizenship activation.
This snapshot of the EU referendum campaigns clearly highlights a continued cleavage between high and low politics (see Figure 1).The under-representation of both women in public debates and women's issues in the material produced by the campaigns demonstrates the continued marginalisation of key constituencies and their concerns.This, in turn, says something fundamental about how activation,

Figure 1 :
Figure 1: 'Vote Leave' versus 'Stronger In' share of mentions in online campaign material

Table 1 :
Gender correlation and policy area in voting at the EU referendum Source: Own calculations.

Table 2 :
Gender correlation and information source about EU referendum Source: Own calculations.

Table 3 :
Gender correlation and discussion about EU referendum Source: Own calculations.

Table 4 :
Logistic regression vote*gender with information sources and discussion of EU referendum Source: Own calculations.

Table 5 :
Logistic regression vote*gender with knowledge of the EU

Table 6 :
'Vote Leave' versus 'StrongerIn' policy codes in online campaign material Source: Own calculations.