You can download the current issue of the Oxford Monitor of Forced Migration here,
or browse the individual articles below.


A Loss for Words: Visual Representations of Migrants and Refugees in the Leave and Trump Campaigns

Updated: Aug 21, 2019

By Theophilus Kwek & Anne Martine Norli Solstad

@DonaldJTrumpJr tweet, 19 September 2016, ‘This image says it all...’. From Kittos, D. Skittles, 15 January 2010, available from:

In 2016, both the ‘Leave’ campaign in the UK and President Trump’s electoral campaign in the US used visual representations of refugees and migrants to great effect in mobilising support for the restrictive immigration policies espoused by their respective platforms. This paper outlines the literature on understanding the role of images in public discourse, and draws on established methodologies to analyse the portrayal of migrants in these two campaigns. It identifies three predominant visual frames that were adopted, and argues that these frames effectively linked debates over immigration policy with other live issues so as to amplify the popular sentiments associated with these grievances. Finally, it reflects on the responsibilities of those who produce images and the possibilities of reframing migration in today’s world.


2016 was marked by two political campaigns in the Anglophone world which drew comparisons for their thematic similarities and unexpected successes at the ballot-box: the cross-party ‘Leave’ campaign in the UK, ahead of the referendum on its membership in the European Union, and Donald Trump’s Republican campaign in the US, ahead of Presidential elections in November. Mainstream commentators were quick to identify these similarities. Even before the referendum, for example, a BBC presenter laid out the reasons ‘Why Brexit Could Signal Trump Winning the White House’ (Kay 2016), and several days after the Leave campaign’s victory, a New Yorker columnist warned that ‘parallels with the Trump campaign could not be more obvious’ (Wood 2016). Immediately after the Leave vote, then-candidate Trump publicly hailed the victory as a precursor to his own (Collinson 2016).

Among other similarities in their messages and personalities (Curtis 2016), immigration stood out as a key theme in both campaigns. Since the outbreak of conflict in Syria, political debate in the US and UK had increasingly revolved around immigration-related issues. Not only had voters come to see these issues as increasingly important (IPSOS Mori 2016, Pew Research Centre 2016), observers also noted a swing towards the ‘new right’ on immigration issues (Ingram 2016:92). The fact that both Trump and Leave campaigns actively promoted restrictive immigration policies offers at least a plausible explanation for their success, given the distinctive similarities in the images accompanying these campaigns.

Across political platforms, images are not only powerful stimuli for engagement; they transcend differences in education and literacy, potentially making information more accessible. The predominance of social media – which privilege visual material – as sites of political messaging suggests that images carried a special salience in the 2016 campaigns. It is therefore important to analyse the Leave and Trump campaigns’ portrayals of refugees and migrants through their respective use of images, to understand how they catered to voters’ preconceptions.

Specifically, images allowed political actors in both campaigns to frame of immigration-related issues to their advantage. Issue framing is the act of focusing on strategically-chosen aspects of reality, making them noticeable and memorable (Entman 1993). In both the Leave and Trump campaigns, visual elements of images were used to craft and emphasise selective justifications for restrictive immigration policies. Our analysis seeks to identify how these perspectives were promoted through visual patterns across both campaigns. In particular, we ask: How were migrants and refugees visually represented in these campaigns? What were the mechanisms associating migrants and refugees to negative narratives? And, as the two campaigns progressed and overlapped, did distinct visual patterns emerge to differentiate them?

We begin by outlining the literature on the role of visual analysis in understanding public discourse, which lays the bedrock for our methodology; at the same time, we reflect on the challenges of approaching and interpreting visual material. We then flesh out three thematic trends in the presentation of migrants across both campaigns, and, in a brief concluding section, discuss their implications and associations. It is our hope that the following exploration contributes to how we engage with, and understand, visual presentations of migrants in the public sphere today.

Reading Images in the Public Sphere

Much of the existing literature on public discourse favours textual modes of political messaging, such as speeches, news editorials, and literary material (Waheed et al 2012, Jacobs and Townsley 2011). Where images are studied, they are often ‘regarded as secondary to texts by many researchers’, and taken to occupy ‘merely a supportive role’ in what is communicated (Banks 2012: 296). There are two reasons, however, why images ought to have a more central place in public discourse analysis: they convey particular representations of public meanings, and do so in ways distinct but inseparable from rationalist modes of public reasoning.

First, images have a unique role in representing meanings in the public sphere. The ‘public sphere’, in Habermas’s conception, is the realm of ‘political confrontation’ distinct from both the private sphere of the home and the governing structure of the state (Habermas 1991: 27). It is where public opinion is contested, and where democratic governments must seek influence, support and legitimacy. In this sphere, opposing views and perspectives can only be made ‘visible and legible’ to others through ‘dynamics of representation’ (Johnson 2011: 1017); as Bleiker puts it, ‘meanings are made public through representation’ (2001: 515). It is this role of representation that images perform when circulated in the public sphere, serving as visual shorthand for the ideas deployed therein, and conveying them in highly unique ways.

One medium of public circulation where images have a distinctively representative role is in photojournalism. Photographs often accompany news headlines to depict situations of significance or crisis, thus allowing readers familiar with the same ‘cultural code’ to know what is described and what response is expected. The images themselves help to ‘express, reinforce, and connote’ that dominant cultural code, thus situating the reader as they engage with her (Banks 2012:4). Moreover, photographs serve to ‘underpin journalistic claims of objectivity’ by creating a sense that ‘the reader [is seeing] what the photographer sees’ (ibid, 296), and the routine use of photography to verify textual meanings suggests that texts are only partially effective in guaranteeing this ‘objectivity’. Even photographic representations, however, present – at best – a contested notion of ‘objectivity’. As visual anthropologists like Poole (1997) have shown, photographs have often been used to entrench race- and class-based ‘canons of taste and distinction’, while ample historical examples demonstrate how earlier forms of imagery have also been deployed to embody political messages in their own right (Anglo 1992: 121, Ryan 1997:13). In these instances, links can be drawn between particular political agents or agendas and images with specific meanings and modes of representation.

In today’s political campaigns, images are predominantly circulated on the internet. As early as 2002, Papacharissi acknowledged that while internet connectivity did not necessarily ‘ensure a more representative and robust public sphere’, online engagement could at least constitute a ‘public space’ across spatial and cultural boundaries (2002: 12). More recent scholarship suggests that the online political environment resembles a ‘dispersed’ public sphere tending towards heterogeneity, as opposed to traditional ‘mass media [which] tend to produce homogeneity’ (Rasmussen 2012: 97-8). In particular, social networking sites support, within such ‘dispersed’ public spheres, a ‘participatory culture’ where ‘members believe their contributions matter, and feel some degree of social connection’ (Mazali 2011: 290). When images are shared across public spaces online, they cross boundaries present in traditional print media, and must be understood in light of heterogeneous discourses that proliferate on the internet. They must also be understood as artefacts meant to engage directly and democratically with viewers, who are often encouraged to ‘like’, share, or comment on them.

The second important feature of images is their power to communicate meanings in non-rationalisable ways. Aesthetic elements lend images dramatic, often emotive effects distinct from more rationalist modes of political messaging. As Johnson argues, images operate ‘at the level of the aesthetic ... partly on the level of perception and emotion rather than thought’. As such, they can ‘go beyond language’ in ‘shaping and suppor[ting] dominant narratives’ (2011: 1017). One effect is that, when paired with textual messages, images disrupt viewers’ assessments of the messages’ validity and salience, and can also determine the extent of the message’s impact on public discourse. Greer observes that by ‘adding a face to a name’ or ‘establishing the identity of key players’, for example, images are able to ‘lend a dramatic or sensational edge to an otherwise “ordinary” piece of news’ (2003: 79). In other words, images do not only invite rational engagement but influence, at a subconscious level, the likelihood and emotional valence of that engagement.

The aesthetic nature of images demands modes of analysis that do not rely on simple notions of ‘what’ is depicted or what ‘message’ is intended. They must be approached in ways that apprehend how pictures, in Barthes’ words, ‘are more imperative than writing’, in the sense that they ‘impose meaning at one stroke, without analysing or diluting it’ (1972:110). One approach, Hansen argues, is to understand images as ‘icons’ or ‘visual nodal points’. When particular images (often attached to a particular moment in political history) become ‘easily recognisable and widely disseminated’, they can ‘activate strong emotional identification or response’, which allows them to circulate further with an almost immediate impact (Hansen 2015:265-271). Though not all images achieve this level of circulation (and hence iconic status), the idea of the icon, derived from religious history, provides a helpful way to understand why those that doare able to speak powerfully to a wide range of audiences. These include the Leave campaign’s infamous ‘Breaking Point’ banner (Stewart and Mason 2016) or the image of Alan Kurdi, a drowned child on the beach (Williams 2016). As Malkki and others have argued, the absence of contextual data in some of these images makes it easier to gloss over the fact that images of refugees are ‘taken of a particular ... situation of migration in a specific location’ (1995:9), and render the images re-usable, universal, and transferable.

Many approaches have been developed for analysing images as ‘data’ (Banks 2014:4). In their 2008 study on centre-right British newspapers, Jones and Wardle suggest that meanings are constructed through: ‘(a) the image itself (its contents, framing, colour, quality, and composition); (b) its relation to other images; and (c) their relation to the headline text’ (2008:60). These elements can be studied on two levels, namely, through content analysis,focusing on the ‘relative size, number, and content of images’, and visual analysis, focusing on the juxtaposition and arrangement of an image in relation to other visual elements (Jones and Wardle 2008:60). Others rely on similar principles, but point out that even so, visual media are ‘open to multiple interpretations’: there is no ‘correct’ reading of why two images may be paired on a broadsheet, for example (Banks 2012:299). We have chosen to adapt elements of their approach (see ‘Methodology’), while remaining conscious of these ambiguities.

Besides recognising the value of studying images as a distinctive component of contemporary political discourse, however, it is also important to recognise the limits of doing so. The aesthetic and emotive aspect of images present one key limitation: that images are ‘simply too messy, too rich, too particular to be reduced to abstraction’ (Banks 2014:16). Regardless of our analytical frameworks, this quality means that images – especially the most powerful ones – will, at some level, defy explanation. Another limitation is that images, in themselves, are only one side of the story. While studying images in context may reveal something of their intended representations, it is far more difficult to measure the receptionof images. ‘Without empirical investigation’, in other words, the researcher can ‘never know how viewers actually respond to the images presented’ (ibid, 8). Similarly, it is impossible to presume the viewer’s positionality, or to measure how much of what the viewer perceives is derived from an image, as opposed to information she already possesses from context or prior knowledge. These obstacles should not discourage the researcher altogether. As Banks again puts it: researchers can be ‘honest’ about their restrictions, and see the ‘apparent limitations of visual methodologies as synonymous with the limits of human self-knowledge itself’ (Banks 2014:16). It is with this honesty that we approach our task.


We first chose to focus on images produced or shared by official platforms behind the Leave and Trump campaigns. However, we quickly realised that both were far from cohesive, but rather a family of related campaigns. In the UK, the two platforms which enjoyed the widest following were and Vote Leave. Whereas Vote Leave was chosen as the official campaign by the Electoral Commission in April 2016, it was Nigel Farage from the campaign who represented the Leave side on ITV’s live debate, opposite Prime Minister David Cameron who represented the Remain side, on 8 June, 2016. In the US, the Trump campaign produced fewer images of immigrants, but often shared third-party news articles and images. We chose to include these, along with videos produced by the Trump campaign, because they formed a significant part of its overall visual output (Brightcom 2016).

We sourced these images from the official campaign websites, and their Twitter, Facebook and Youtube profiles, accounting for posts in the one-year period before the EU referendum on 23 June, 2016, and the Presidential election on 9 November, 2016 respectively. Instead of scrolling through these posts manually, we used systematic search strings (key phrases such as ‘migrant crisis’ or ‘refugee deal’) to identify the images and their accompanying captions. As Banks observed (2012:6), textual captions are an essential part of the framing process. Conversely, manual selection might have biased us towards the most recent or visually provocative examples, and using search strings minimised the risk of a skewed sample. Since the legal statuses of ‘refugee’ and ‘migrant’ were often used interchangeably, our search strings included both terms; few discernible visual differences could be drawn between images paired with each term.

While we originally hoped to prioritise images produced by the campaigns themselves, we recognised that from the ordinary viewer’s perspective, these did not stand out from third-party news articles or websites shared by these campaigns. This is largely due to the layout of social media platforms, which allow users to frame the interpretation of third-party images within larger campaign messages by pairing them with customised captions. When an article is shared on Facebook, for instance, the layout privileges the image and headline from the original article but displays little else, giving the user considerable leeway with the accompanying caption. From the viewer’s perspective, the images and captions are inseparable, and sharing ‘header’ images from news articles allows users to exploit their visual impact. Given these possibilities and strategies, campaign platforms can be more accurately described as ‘curators’, rather than merely ‘producers’ of visual material.

Having selected our images, we applied a loose adaptation of Jones and Wardle’s (2008) approach to image analysis. We chose to focus especially on the components of the images themselves, along with their relationship to the headline text, and their overall effect. Specifically, we paid most attention to their compositions, symbolic elements, the demographics of their subjects, and any in-frame text or infographics used: components which have also received greatest attention from other scholars (Bleiker et al 2013). Where symbolic elements were analysed, we inferred their referents and connotations from the perspective of Anglophone viewers of voting age, broadly familiar with elements of public imagery and Western popular culture.

As a preliminary example, Figure 1 – shared from Donald Trump Jr’s Twitter account, and bearing the Trump campaign logo – compares the US’s ‘Syrian refugee problem’ to a bowl of coloured sweets, with the caption ‘If I had a bowl of skittles[sic]and I told you just three would kill you. Would you take a handful?’ The collective terms (‘bowl’, ‘three’, ‘handful’) place the numerical aspect of the refugee situation at the visual focus of the image, while the dramatic language (‘kill’) immediately invokes security concerns. The image depersonalises refugees, and implies that one might not be able to tell which of the Skittles are poisonous, analogous to the idea that one might not be able to tell which Syrians are really terrorists in disguise.

Figure 1: @DonaldJTrumpJr tweet, 19 September 2016, ‘This image says it all...’

There are, of course, risks of reading images in this way: all ‘actions, objects, and practices are socially meaningful’, and these meanings are shaped by highly specific contexts (Fischer 2003: 73). Likewise, meanings derived from analysing images are inseparable from analysts’ contexts and identities, and other researchers have taken pains to acknowledge their own positionalities vis-à-vis their chosen subject (c.f. Banks 2012: 298-9). Both authors of this paper were resident in the UK throughout 2016, and part of the voting population at which some of these images were aimed. However, we were (and are) neither British nor American by nationality, and came from cultural contexts which were not strictly Anglophone, and thus could be seen as ‘outsiders’ to mainstream politics in both countries. Our views should be read accordingly.

Sinner, Stranger, Scrounger: Three visual frames

Since the 1951 Refugee Convention established the legal definition of a ‘refugee’, Johnson (2011: 1016) has identified a re-framing of the ‘image of the refugee … from a heroic, political individual to a nameless flood of poverty-stricken women and children’ (also Bleiker et al 2014:193). Many contemporary media portrayals render refugees as helpless and vulnerable, borrowing visual tropes of the ‘Madonna and Child’ (Wright 2014:462), while those fleeing the effects of environmental degradation have also been depicted as such (Mehtman 2014:422). We found that representations of refugees and migrants in the Leave and Trump campaigns differed from these narratives: though a numerical element remained important, the ‘nameless flood’ consisted of racialised images of young, dark-skinned men who were neither heroic nor vulnerable.

Emphasising the numbers of migrants is a key technique of anonymisation. Refugees are often portrayed as ‘shadowy strangers’, or as persons with ‘no visible features’ (Banks 2012:10). At greater distances and group sizes, individuals are less visible, and images are not ‘likely to create compassion and empathy in viewers’ (Bleiker et al 2014:193). Other studies have shown that images of migrants tend to include groups of over fifteen persons (Bleiker et al 2013: 404), and our image sample overwhelmingly followed this trend.

Beyond the numbers, however, refugees and migrants are frequently also portrayed as both deviants and strangers. As Banks (2012: 293) has shown, the migrant is often coded as a ‘criminal immigrant’, a visual framing which fuses ‘the otherness of the stranger with the otherness of the deviant’. This resonated with our findings, where migrants were presented by both campaigns as morally deviant, and as strangers who were alien to the social majority. While the former relied on a perception of their behaviour as irregular, and thus threatening, the latter relied on their being seen as culturally incompatible, and both dynamics were amplified through images that stressed the gendered and racialised dimensions of their ‘otherness’ vis-à-vis a dominant ‘white’ majority. As Sanchez (2018) and others have shown, this reflects a wider trend among contemporary presentations of migration flows to the US that focus on the ‘journeys of gendered and racialised Others’, and portray them as ‘invasions’.

In addition, a third theme emerged among the images we studied: that of migrants as burdens on the state. This portrayal did not fit easily into the sinner-stranger dichotomy, as the focus was not on how migrants were ‘other’ to the host country and its inhabitants, but on how similarthey were, as intimidating competitors for work and social provision within the polity of the host state. These three themes – the presentation of refugees and migrants as threats to security, disruptions to society, and burdens on the state – are explored here in turn.

The migrant as deviant

First, images from the Leave and Trump campaigns constructed immigrants as ‘deviants’, and hence as threats to public and personal security. These images were able to draw on established ideas of the state’s responsibility for its citizens’ safety, which had special resonance with conservative political audiences in the US (Meese 2011) and UK (Dathan 2016).

First, migrants and refugees were presented as threats to publicsecurity. The images in our sample drew on contemporary, widely-held concerns generated by terrorist attacks in the US and Europe, including false stereotypes associating the attacks with refugees (Funk and Parkes 2016). They implied (and occasionally stated) that the EU’s current immigration arrangements and the Clinton campaign’s push for more liberal immigration policies were not only irresponsible, but detrimental to the security of the UK and US. The terrorism narrative was most visible in images shared from Donald Trump’s own Facebook page during the campaign, with illustrations of ISIL fighters in almost every piece about migration (see Figure 2b).

Figure 2a, left: website, 23 March 2016, ‘Now is the time…’). Figure 2b, right: Donald J. Trump’s Facebook post, 12 February 2016, ‘Trump warns...’

In Figure 2a, three armed security personnel have their backs turned to the camera, and are facing a large crowd of young men with urgent expressions. Although none of the UK’s borders corresponds to the scene depicted, the accompanying headline (‘Now is the time…’) suggests that this is intended to represent a European land border. The presence of a small number of security personnel implies both that it may be necessary to deploy force at the border, but also that the existing deployment is insufficient by far to prevent the crowd from crossing an invisible border to where we, the viewers, are. This subtle connotation of a military threat is rendered explicit in Figure 2b, where the migrant is directly coded as a masked militant. Over the last two decades, the figure of the masked militant has become heavily associated with presentations of Muslims in the UK’s news media (Moore et al 2008) as well as news and film in the US (Ramji 2003). Indeed, the ways in which these images are deployed are reminiscent of Baker et al’s longer-term findings (2013) of how Muslim men have been constructed in the press as ‘passive and active agents’ in the threat posed by religious violence to British society.

Nevertheless, not just public security was at stake. Immigrants were also presented as threats to personalsecurity, through portrayals of them as morally unscrupulous or deviant. Many images accompanying articles shared by the Trump campaign included handcuffs and other police effects (see Figure 3b). In the Trump campaign videos, too, migrants were pictured with blurred or hidden features, evoking forbidding notions of subversion and racialised criminality.

Figure 3a, left: website news article, 9 September 2015, ‘Ten percent of refugees…’ Figure 3b, right: Donald J. Trump Facebook post, 2 August 2016, ‘Police arrest…’

Figure 3a, shared by, depicts policemen apprehending a ‘delinquent’ who (in accordance with the headline) is assumed to be a refugee. Even though prominent supporters of the Leave campaign, such as Security Minister John Hayes, blamed the EU’s centralised bureaucracy rather than immigrants for reduced security (Dominiczak and Hope 2016), images like these still focused on the criminality of refugees and the potential threat they represented to everyday life. Given the British context, the term ‘delinquent’ was poised to evoke irrational deviance among youths, gesturing towards the moral panics of an earlier era surrounding ‘hooligan’ subcultures (Cohen 2011), which took on further class- and race-based undertones in media coverage of the August 2011 riots, which followed the police shooting of Mark Duggan (LSE and The Guardian 2011).

The migrant as alien

A second common theme was the presentation of migrants as alien, and hence disruptive, to the culture of the host community. Other scholars have examined dynamics of ‘othering’ within anti-migrant discourse in the UK and other European contexts (see for example Strani and Szczepaniak-Kozak 2018). Often, various identity markers are conflated or essentialised, and given normative relevance in political debate over immigration issues (Rettberg and Gajjala 2016). The Trump and Brexit campaigns were no different. Conflated identity attributes – such as culture, ethnicity, religion, age and gender – were invoked through images (and their accompanying texts) to associate disruptive or unruly attributes with ‘otherness’. These were powerfully juxtaposed, in turn, against ideas of ‘European’ civility and order.

Fig. 4a, left: on Twitter, 3 April 2016, ‘Greece on brink...’ Fig.4b, right: Donald J Trump on Facebook, 30 August 2016, ‘One million Muslim...’

The images above focus on migrants’ numbers, age, and gender, but also their ethnicities, which suggest Middle Eastern or North African heritage. Refugees are often racialised (Mehtmann 2014), and, as the caption of Figure 4b suggests, the stereotypes associated with these ethnicities are conflated with those surrounding Islam. Not only does the caption present a racialised portrayal of ‘immigrants from the Muslim world’; the image itself implies that the ‘Muslim world’ is about to penetrate – and overwhelm – the US, which is only defended by flimsy tape. These images invoke a stereotype of the ‘Middle-Eastern man’ as threatening and dangerous, especially for women (Rettberg and Gajjala 2016). There is also a thinly-veiled fear of demographic change; that a greater proportion of the population will be dark-skinned.

Presentations of ‘otherness’ are doubly stark when contrasted with images of ‘order’. Scepticism against immigrants is often justified through references to cultural incompatibility, and implications that a fragile cultural balance, or prevailing expectations of social etiquette, will be upset by the entry of migrants. This theme was more prevalent in the UK context, perhaps rooted in cultural preferences or general public discourse on the importance of rules, queues, and tidiness. In many images, exemplified below, we see migrants’ tents filling up the frame, in the vicinity of sleeping bags, plastic bags, and belongings scattered in bushes and on the ground. Without providing any context explaining the ‘mess’, the image conflates migrants with their surroundings, implying that this messy situation might be transferred to the UK.

Fig. 5a, left: on Twitter, 8 February 2016, ‘Is it really wise...’ Fig. 5b, right: on Twitter, 26 November 2015, ‘EU can’t take...’

More urgently, in Figure 5b, which shows young men pushing to get on a bus, the appeal is towards ‘European’ values of queuing and orderliness, which are seen to be lacking among migrants. Disorder is thus presented as an essentialised attribute of migrants, which lends itself to invoking moral panic; by ‘amplifying deviance’, the media encourages ‘increased public anxiety and intensified policing’ (Banks 2012:2). As Esses et al have argued, the perception that a particular group ‘lacks prosocial values’ can form the basis of a view that the group is ‘less human and thus less worthy of humane treatment’ (2013:521).

The migrant as burden

The third theme, seeing migrants as burdens to the state, departs from the ‘stranger/deviant’ duality to frame migrants as having similar needs and hence competing for the same resources. To make this case, images focused on the numbersof immigrants, rather than specific challenges to state provision. Both campaigns framed a rapid increase in immigration as exceeding state capacity; while the Leave campaign depicted migrants as posing an increased demand for welfare benefits, both campaigns accused migrants as creating competition for work.

Concerns about welfare were predominant across Leave platforms, and less so in the Trump campaign. Figure 6a shows a queue of immigrants positioned across the image to resemble a river, recalling then-Prime Minister David Cameron’s controversial description of immigrants as a ‘swarm’, or the Daily Mail’s reference to a ‘tidal wave’ (Shariatmadari 2015). The depth of field created by the image’s blurred backdrop suggests that the line of immigrants extends indefinitely, and the focus is not on any individual but on the number of people. A bright red slogan, ‘Breaking Point’, implies that a limit has been breached, while the subtitles imply that the state would only fulfil its mandate by imposing restrictions on this number.

Figure 6a, left: tweet, 16 June 2016, ‘Breaking point...’ Figure 6b, right: VoteLeave website graphics, April 2016,‘EU immigration is…’

More specific images create perceptions of an increased (and unsustainable) demand for welfare provisions. In Figure 6b, a statistic about rising immigration is paired with the hashtag ‘#TakeControl’, and the slogan ‘Save our NHS’. The image depicts figures in red – matching the red text of ‘EU immigration is growing’ and the icon of a red EU flag – who outnumber figures in yellow, meant to depict existing citizens. The implication is that a group of people flying the EU flag are eroding the stable pyramid-shaped structure of the NHS, creating a ‘big strain’. The comparison in the caption to ‘a city like Newcastle’ further draws on public conceptions about the Northeast’s structural lack of resources, and targets an audience that is far more likely to identify with the population of Newcastle than with new immigrants.

Visual patterns focusing on work were found in both the Leave and Trump campaigns. The images largely portrayed migrants as competitors to existing citizens of the UK and the US, and focused on their ability, rather than their vulnerability. People presented in the images were often of working age, male, and able-bodied. Migrants were presented as competing for the same jobs that low-skilled workers currently held. In some images from the Leave campaign (e.g. Figure 7b), ethnic differences were in fact downplayed: images of ‘white’, European migrants were chosen to suggest that a majority of migrants were in fact similar to, and hence competing against, British workers for jobs.

Figure 7a, left: DonaldJTrump Youtube video, 21 September 2016, ‘America’s veterans...’, Figure 7b, right: tweet, 4 August 2015, ‘EU pays jobless migrants...'

In the US too, a similar narrative could be identified. One Trump campaign video (a still of which is shown in Figure 7a) claimed that ‘Large corporations bring in many thousands of low wage workers…from overseas and across the border to fill jobs that could easily be filled by our veterans’. Here, the migrants are presented as shadows and shapes, indistinguishable from other workers – drawing on the pre-existing concerns of many working-class voters about being mere digits in an impersonal economy, one that caters to the interests of ‘large corporations’. While Republicans from an earlier era may well have wholeheartedly supported such corporate interests, this language is intended to tap into the anxieties of a post-industrial demographic – for whom transnational corporate powers are more likely to be seen as a threat to social cohesiveness (at least, for the cultural majority) and to national sovereignty (Brown 2010). Moreover, for many on the political right, veterans embody a republican ideal of ‘deservingness’ within the state, having earned their access to public provision through service and sacrifice (Cowen 2006). In the perceived zero-sum game of economic competition, there is no question of whether they, or low-wage migrants, are more deserving of American jobs.


In the preceding pages, we have presented a working methodology of analysing campaign images on social media platforms, taking into account the ease and strategic value of pairing particular captions or messages with images from otherwise credible news sites. Based on this approach, we have examined images used in both the Leave and Trump campaigns, and assessed how they created and reinforced perceptions of refugees and migrants as threats to security (through difference and deviance), disruptions to society (through ‘otherness’ and disorderliness), as well as burdens on the state (by competing for work and welfare). Each of these frames tapped deliberately into popular anxieties that were already salient in the political discourse of the UK and US respectively, bringing them together in the figure of the racialised, gendered migrant, and mobilising images strategically to leverage on their combined electoral appeal.

These findings are, inevitably, inexhaustive. Given our methodology, it is difficult to tell which of the three frames examined above had greatest salience, and with which audiences. More extensive investigation (as well as real-time data on the evolving conversations surrounding both campaigns) would be required to understand the precise ways in which images not only amplified pre-existing concerns, but shifted perceptions and seeded new misconceptions. On a more personal level, having examined the campaign material in retrospect, it has been difficult for us as researchers to dissociate our social media artefacts from before the UK referendum or US election with the political facts of anti-migrant policy and discourse that took shape after both votes. Nevertheless, the findings provide some basis for critical reflection.

The extensive use of images from external websites and news articles – images not produced by the campaigns themselves – reminds us that circulation in the public sphere implies heavy responsibilities for all producers of visual material. Technology allows the easy co-optation of public images into explicit political discourses; even ‘real’ images, in other words, can be used to entrench false (and harmful) narratives (see Esses et al 2013). Since the 2016 campaigns, governments around the world have placed increased emphasis on legislating safeguards against the spread of deliberate online falsehoods. Few, however, emphasise the potential impacts of images, or the responsibilities of those who create them. Particularly where existing power differences allow the subjects of such representations little right of reply, visual producers must be especially careful of the terms and frames they employ (Cole 2016).

At the same time, concerns about security threats, societal changes, and insufficient state welfare services are not new. Refugees and migrants have arguably become such salient visual subjects in both campaigns only because the theme of immigration has proven exceptionally and consistently powerful in amplifying these pre-existing political issues. Political actors with a broadly conservative policy agenda have, in other words, managed the representations of migrants so as to mobilise latent grievances not directly related to immigration. The injustice perpetuated in turning those who are often already in need of protection from political violence elsewhere into subjects of domestic political manipulation should outrage us (Blitz 2017). However, there are also reasons to hope. If the deeper concerns in these campaigns and their treatments of migrants are notabout immigration per se, perhaps there are opportunities to resist the narrative they perpetuate: by addressing those underlying grievances at the level of state and community, by adopting positive frames towards immigration issues in wider public discourse, and by ensuring the fair representation of migrant and refugee communities in the production and gatekeeping of this discourse.

It is worth mentioning, in closing, one ethical concern of this study. By using and perusing the campaign images, we too have become part of their audience – and by increasing internet traffic to the relevant sites and accounts, inadvertently raised their viewership and search profiles. As Dauphinée (2007) points out, there are ethical complications involved in using and reproducing any depiction of suffering, and the same may be said of representations that distort and malign refugees and migrants, as those we have used above. Though we have taken pains to exercise critical awareness in our engagements with these images, and urge all our readers to do the same, a deeper point nonetheless remains: we are all implicated in what we see.

Theophilus Kwek is a writer and editor based in Singapore. He has presented his research on refugee issues at conferences in Frankfurt, London, and Singapore, and contributed to platforms such as The Diplomat, Singapore Policy Journal, South China Morning Post, and Refugees Deeply. His most recent research studies the conscription of second-generation migrants in Singapore, reflecting on the norms of conscription and citizenship. He holds an MSc in Refugee and Forced Migration Studies from Oxford University, and works in the healthcare sector.

Martine Solstad holds a BA in Politics and International Relations from University of Cambridge and an MSc in Refugee and Forced Migration Studies from Oxford University.


ANGLO, S. (1992) Images of Tudor Kingship, Guildford, Seaby.

BAKER, P., GABRIELATOS, C., and MCENERY, T. (2013) Discourse Analysis and Media Attitudes: The Representation of Islam in the British Press, Cambridge, Cambridge University Press.

BANKS, M. (2014) ‘Analysing images’, 394-408 in Flick, U. (ed.), The SAGE Handbook of qualitative data analysis,London, SAGE Publications.

BANKS, J. (2012) ‘Unmasking deviance: the visual construction of asylum seekers and refugees’, Critical Criminology: International Journal20(3): 293-310.

BARTHES, R. (1972). Mythologies, New York, Hill and Wang.

BLEIKER, R. (2001) ‘The aesthetic turn in international political theory’, Millennium30(3): 509-533.

BLEIKER, R., CAMPBELL, D., HUTCHISON, E. and NICHOLSON, X. (2013) ‘The Visual dehumanisation of refugees’.Australian Journal of Political Science48(4): 398-416.

BLEIKER, R., CAMPBELL, D. and HUTCHISON, E. (2014) ‘Visual Cultures of Inhospitality’. Peace Review26(2): 192-200.

BLITZ, B.K., (2017) ‘Persecution and the Threat to the Refugee System’. openDemocracy(online). Available from <>(Created 27 January 2017, accessed 1 March 2019).

BROWN, W. (2010) Walled States, Waning Sovereignty, New York, Zone Books.

COHEN, S. (2011) Folk devils and moral panics, London, Routledge.

COLE, P. (2016) ‘Framing the Refugee’. E-International Relations (online). Available from <> (Created 25 November 2016, accessed 1 March 2019).

COLLINSON, S. (2016) ‘Trump on Brexit: America is Next’. CNN(online). Available from <> (Created on 25 June 2016, accessed 1 March 2019).

COWEN, D. (2006) ‘Fighting for “Freedom”: The End of Conscription in the US and the Neoliberal Project of Citizenship.’ Citizenship Studies10, No. 2: 167-183.

CURTIS, J. (2016) ‘US election 2016: The Trump-Brexit voter revolt’. BBC(online). Available from <> (Created 11 November 2016, accessed 1 March 2019).

DATHAN, M. (2016). ‘Britain’s border security must be tightened to keep us safe from terrorists’. Daily Mail(online).Available from <> (Created 20 April 2016, accessed 1 March 2019).

DAUPHINÉE, E. (2007) ‘The politics of the body in pain: Reading the ethics of imagery’. Security Dialogue,38(2): 139-155.

DOMINICZAK, P. and HOPE, C. (2016) ‘Security minister says Britain better placed to fight terrorists outside EU’. Telegraph(online). Available from <> (Created 10 June 2016, accessed 1 March 2019).

ENTMAN, M. R. (1993) ‘Framing: Towards Clarification of a fractured paradigm’.Journal of Communication 43(4): 51-58.

ESSES, V. M., MEDIANU, S., and LAWSON, A. S. (2013) ‘Uncertainty, Threat, and the Role of the Media in Promoting the Dehumanization of Immigrants and Refugees’. Journal Of Social Issues 69(3): 518-536.

FISCHER, F. (2003) Reframing Public Policy: Discursive Politics and Deliberative Practices, Oxford, Oxford University Press.

FUNK, M. and PARKES, R. (2016) ‘Issue Alert: Refugees versus terrorists’. EU Institute for Security Studies (online). Available from <> (Created January 2016, accessed 1 March 2019).

KAY, K. (2016) ‘Five Reasons Why Brexit Could Signal Trump Winning The White House’. BBC(online). Available from <> (Created 20 June 2016, accessed 1 March 2019).

GREER, C. (2003) Sex crime and the media: Sex offending and the press in a divided society, Cullompton, Willan.

HAHN, J. (2016) ‘Immigration Expert: U.S. to Resettle Nearly One Million Muslim Migrants Under One Term of Clinton Presidency’. Breitbart (online). Available from <> (Created 26 August 2016, accessed 1 March 2019].

HABERMAS, J. (1962) The Structural Transformation of the Public Sphere, translated from German by T. Burger (1991), Cambridge, MIT Press.

HANSEN, L (2015) ‘How Images Make World Politics: International icons and the case of Abu Ghraib’. Review of International Studies41(2): 263-288.

INGRAM, A. (2017) ‘Geopolitical events and fascist machines: Trump, Brexit and the deterritorialisation of the West’. Political Geography 57: 91-93.

IPSOS MORI. (2016) ‘Immigration is now the top issue for voters in the EU referendum’. Ipsos MORI Political Monitor(online). Available from < June 2016, accessed 1 March 2019).

JACOBS, R. and TOWNSLEY, E. (2011)The Space of Opinion: Media Intellectuals and the Public Sphere, Oxford, Oxford University Press.

JOHNSON, H. L. (2011) ‘Click to Donate: visual images, constructing victims and imagining the female refugee’. Third World Quarterly 32(6): 1015-1037.

JONES, P. and WARDLE, C. (2008) ‘No Emotion, No Sympathy: The visual construction of Maxine Carr’. Crime Media Culture4(1): 53-71.

LONDON SCHOOL OF ECONOMICS AND POLITICAL SCIENCE and THE GUARDIAN. (2011). Reading the Riots: Investigating England’s Summer of Disorder. Public report. Available from <> (Accessed 4 July 2019).

MALKKI, L. (1995) Purity and Exile: Violence, Memory and National Cosmology Among Hutu Refugees in Tanzania, Chicago, Chicago University Press.

MANDELBAUM, E. (2016) ‘The role of video advertising in the 2016 election’. BRIGHTCOM (online). Available from <>(Created on 26 June 2016, accessed 1 March 2019).

MAZALI, T. (2011) ‘Social Media as a New Public Sphere’.Leonardo44(3): 290-291.

MEESE, E, III. (2011) ‘Who is Responsible for America’s Security?’ The Heritage Foundation(online). Available from 19 August 2011, accessed 1 March 2019).

METHMANN, C. (2014) ‘Visualizing Climate-Refugees: Race, Vulnerability, and Resilience in Global Liberal Politics’.International Political Sociology 8(4): 416-435.

MOORE, K., MASON, P. and LEWIS, J. (2008) The Representation of British Muslims in the National Print News Media 2000-2008, Cardiff, Cardiff School of Journalism, Media, and Cultural Studies.

PAPACHARISSI, Z. (2002) ‘The Virtual Sphere: The internet as a public sphere’. New Media and Society4(1): 9-27.

PEW RESEARCH CENTER (2016) ‘Top voting issues in 2016 election’. 2016 Campaign: Strong interest, widespread dissatisfaction(online).Available from <> (Accessed 1 March 2019).

POOLE, D. (1997) Vision, Race, and Modernity: A Visual Economy of the Andean Image World, Princeton, Princeton University Press.

RAMJI, R. (2003) ‘Representations of Islam in American News and Film: Becoming the Other’, 65-72 in MITCHELL, J., and MARRIAGE, S., (eds.) Mediating Religion: Studies in Media, Religion and Culture, London, A&C Black.

RASMUSSEN, T. (2012) ‘Internet-based media, Europe and the political public sphere’. Media, Culture and Society35(1): 97-104.

REICHLEY J.A. (2010) Conservatives in an Age of Change. The Nixon and Ford Administrations,Washington, D.C., Brookings Institution.

RETTBERG, J. W. and GAJJALA, R. (2016) ‘Terrorists or cowards: negative portrayals of male Syrian refugees in social media’,Feminist Media Studies16(1): 178-181.

ROSE, G. (2006) Visual methodologies: an introduction to researching with visual materials, London, SAGE publications, 2nd ed.

RYAN, J. (1997) Picturing Empire: Photography and the visualization of the British Empire, London, Reaktion Books.

SANCHEZ, G. (2018) ‘Portrait of a Human Smuggler: Race, Class, and Gender Among Facilitators of Irregular Migration on the US – Mexico Border’, 29-42 in: Bosworth M, Parmar A and Vazquez Y, eds. Race, Criminal Justice, and Migration Control, Oxford, Oxford University Press.

SHARIATMADARI, D. (2015) ‘Swarms, floods and marauders: the toxic metaphors of the migration debate’. The Guardian(online). Available from <> (Created 10 August 2015, accessed 1 March 2019).

STEWARD, H. and MASON, R. (2016) ‘Nigel Farage’s anti-migrant poster reported to police’.The Guardian(online). Available from <> (Created 16 June 2016, accessed 1 March 2019).

STRANI, K., and SZCZEPANIAK-KOZAK, A. (2018) ‘Strategies of Othering through Discursive Practices: Examples from the UK and Poland’. Lodz Papers 14(1): 163-179.

WAHEED, M., SCHUCK, A.R.T., NEIJENS P.C., and VREESE, C.H. (2012). ‘Values in the News: Comparing the news coverage of political speeches in developed and developing countries’. Journalism Studies14(4): 1-17.

WEINREICH, H., OBENDORF, H., HERDER, E., and MAYER, M. (2008) ‘Not Quite the Average: An Empirical Study of Web Use’. ACM Transactions on the Web2(1): 1-31.

WILLIAMS, C. (2016) ‘Police arrest 900 Muslim migrants in England and Wales for “sickening” crimes’. Jihad Watch (online). Available from <>(Created 1 August 2016, accessed 1 March 2019).

WILLIAMS, R. (2016) ‘A year after Aylan Kurdi was pictured washed up on the beach, what has changed for Syrian children?’ New Statesman (online). Available from 2 September 2016, accessed 1 March 2019).

WOOD, J. (2016) ‘On Brexit, the 2016 Euro Championship, and Trump’. The New Yorker (online). Available from <> (Created 27 June 2016, accessed 1 March 2019).

WRIGHT, T. (2002) ‘Moving images: The media representation of refugees’. Visual Studies17(1): 53-66.

WRIGHT, T. (2014) ‘The Media and Representations of Refugees and Other Forced Migrants’, 460-472 in E. FIDDIAN-QASMIYEH, G. LOESCHER, K. LONG, and N. SIGONA, (eds.) The Oxford Handbook of Refugee and Forced Migration Studies, Oxford, Oxford University Press.

©2019 by OxMo