Introduction

Europe is facing an unprecedented migration and refugee crisis as a result of the ongoing civil war in Syria, which began in 2011 and has resulted in over 250,000 casualties as well as millions of refugees and internally displaced persons. According to the International Organisation for Migration (2015), more than one million irregular migrants and refugees crossed into Europe in 2015. This has given rise to a series of issues, including those involving EU security, Europe’s common values and destiny, the readiness and policy planning of the Syriza-led government in Greece, and the willingness of Turkey to cooperate with the EU.

Some media outlets soon began to write about ‘the largest movement of people that Europe has seen since 1945’ (Financial Times 2015). Others, adopting a more critical approach, stressed the need for concerted EU action (Barakat and Zyck 2015). Indeed, since 2011 several humanitarian organisations have been sounding the alarm about the growing refugee crisis. However, the wake-up call in public opinion did not come until four years later, when the shocking images of the drowned three-year-old Syrian refugee Aylan Kurdi emerged. This sharp increase in interest from the public and the media was one of the main factors that pushed the EU to pursue an agreement with Turkey that would aspire to end the irregular migration from Turkey to the EU.

This article aims to explore the ways in which the migrant and refugee crisis has been covered by online media in Germany, Greece and the UK. More concretely, the article aims to both identify the way the sampled online newspapers in the above-mentioned countries portray the migrants and refugees, and provide a preliminary analysis of the aspects on which journalists focused when covering this crisis. In this regard, this study does not constitute a large-scale analysis of the media discourse on the refugee crisis as a whole.

Prior efforts to investigate the media discourse on immigrants and refugees

The media holds a pivotal role in disseminating information, spreading knowledge and shaping ideologies, as well as exerting influence over societies. In this regard, the media discourse on immigrants and refugees has attracted considerable attention from several scholars. Focusing on discursive mechanisms of representing and interpreting the ‘others’ (i.e. immigrants and refugees), Van Dijk (2000) argued that the media contribute to the reproduction of stereotypes, prejudices and eventually racism. In a seminal paper that employed a critical discourse analysis of two Australian newspapers, Teo (2000) unravelled how criminal-related stereotypes for Asian immigrants had been systematically reproduced. Moreover, Santa Ana (1999), focusing on metaphors in the discourse of the Los Angeles Times, noticed that immigrants were presented as animals, debased persons, weeds or disease, while Erjavec (2003), when identifying the media discourses of immigration in Slovenia, stressed that they promoted ‘moral panic’, were highly concerned for national security and were hostile towards foreigners. More concretely:

In the construction of moral panics, the journalists gave up their right to take reporting as a process of searching for the truth and the story was told by the dominant ideology, instead of by them. As loyal followers of the professional ideology, they objectified the adopted xenophobic discourse of the dominant ideology by reference to the creators of such discourse (Erjavec 2003, 99).

In the same vein, while Leudar et al. (2008) conducted interviews with refugees/asylum seekers in the UK, they simultaneously examined the way that the British broadsheets presented refugees/asylum seekers, and identified ‘hostility themes’ in the newspapers’ treatment of them. These themes were reflected in media coverage and local inhabitants’ social expressions (Leudar et al. 2008). Finally, Gabrielatos and Baker (2008) documented negative representations of migrants within the British press. More concretely, among the most frequently used verbs in their sample were ‘fleeing, sneaking, and flooding’ (Baker et al. 2008). In contrast, during the Balkan conflict of 1999, the newspapers in the UK seemed supportive of affected/displaced people, adopting positive discursive practices, although some stereotypes and prejudices were reproduced (Khosravinik 2009).

Methodology

Two mainstream online newspapers from each of the above-mentioned countries were selected, and a combination of the methodologies associated with corpus linguistics (CL) and discourse analysis was employed (see Table 1). First we examined the wording of the articles in detail by enumerating the words (collocates) that were most likely to appear in the text either before or after the following keywords: (i) migrants, (ii) immigrants, (iii) refugees, and (iv) asylum seekers. In a similar study, examining the discourse on refugees and asylum seekers in the UK press, Baker et al. (2008, 277) indicate: ‘the CL analysis started with the examination of relative frequencies and emerging statistically significant lexical patterns in the corpus and sub-corpora mainly involving the four terms in focus: refugee(s), asylum seeker(s), immigrant(s), migrant(s)’.

Table 1 L1 and c-collocates
Full size table

The following collocates were employed:

  • Those words placed immediately to the left of the above-mentioned keywords or descriptors, labelled ‘L1 collocates’.

  • Those words that regularly emerged within 10 words of our keywords, either to the left or to the right, labelled ‘consistent collocates’ (c-collocates) (see Baker et al. 2008; Allen and Blinder 2013).

In this regard, we endeavoured to collate the vocabulary that the various online newspapers used to portray the migrants and refugees, as well as to highlight the aspects on which the selected online media focused when covering this crisis. Our research is based on an analysis of a corpus of 1340 articles published online between 20 March and 31 May 2016. In contrast to previous studies that covered several years of research and a large amount of data (e.g. Baker et al. 2008; Allen and Blinder 2013), we did not use CL software. Rather, we manually documented the words that appeared in conjunction with our keywords. Having adopted this approach, we were able to identify and interpret the meaning of the phrases that included the descriptors and the c-collocates as well as the relevant sentences. According to Allen and Blinder (2013), computer-assisted methods are able to analyse a large amount of data, eliminating the possibility of human error. However, the authors would argue that the current study did not require a computerised method since the period of research was so short (approximately two months). This period went from 20 March 2016 to 31 May 2016.Footnote 1 With regard to the selection of 10 words either side of our keywords (‘c-collocates’), this number was partly decided on arbitrarily and partly in such a way that several words could be captured for meaningful analysis.

However, CL studies seem unable to provide a qualitative analysis of the linguistic context (e.g. tone and meaning) where collocates are used within sentences or phrases. Therefore, as a second step, we tried to further present and interpret the main aspects of the top collocates of our target words (see Table 3). For instance, what was written about ‘Idomeni’, ‘hotspots’ or ‘Turkey’?

As was noted above, this research takes the EU–Turkey deal of 20 March 2016 as a landmark. The selection of online broadsheets was based both on ease of research and the authors’ consideration that the content and discourse of the online versions of the mainstream media are to a great extent consistent with the print and broadcast media. In addition, online journalism is progressively gaining a significant audience.

The online newspapers were selected in accordance with their political affiliation—centre–right/centre–left, as well as their circulation (see Table 2). From Germany, we selected Die Welt and Süddeutsche Zeitung, while from Greece Kathimerini and To vima. Last but not least, from the UK we examined The Guardian and The Telegraph.

Table 2 Categorisation of the online broadsheets in accordance with their political affiliation
Full size table

Regarding the countries that serve as case studies, the selection of Germany, Greece and the UK serves the purpose of covering as many EU member states as possible and representing countries from the northern and the southern part of the EU. Additionally, Greece seems to have been the EU country most affected by the refugee crisis, especially since the closure of the ‘Balkan route’, as it is playing host to tens of thousands of refugees and irregular migrants, according to the latest reports. A humanitarian crisis has been sparked in Greece due to its geopolitical position, primarily its border with Turkey, along which smugglers continue to operate. With regard to Germany, it was agreed that it would be interesting to capture how the online coverage from the German newspapers presents and comments on the crisis, taking into consideration the decisiveness and efforts of German Chancellor Angela Merkel to implement the quota plan to deal with the refugee crisis.

Another important element regarding the selection of the aforementioned countries is that extreme right-wing populist parties have gained significant support within both Germany (Alternative für Deutschland) and Greece (Golden Dawn), and there has been a notable rise in Euroscepticism in these countries.

As far as the UK is concerned, taking into consideration its sceptical position within the EU as well as the result of the EU referendum, it was agreed that it would be interesting to capture and analyse the UK newspapers’ perspectives of the refugee crisis. Unfortunately, due to language and space limitations, the eastern part of the EU was not covered.

Findings and discussion

L1 collocates

Table 3 shows the words (L1 collocates) that, in the articles examined, preceded the target words of ‘migrants’, ‘immigrants’, ‘refugees’ and ‘asylum seekers’. L1 collocates are the words used to describe or define these words.

Table 3 L1 collocates of ‘migrants’, ‘immigrants’, ‘refugees’ and ‘asylum seekers’ (top 30)
Full size table

To begin with, as one might expect, the most common collocate in all three countries of our study was not a word but a number (in 62.9% of cases in Greece, 36.1% in Germany and 71% in the UK), specifying the number of immigrants/refugees/asylum seekers arriving/being deported/being repatriated to/from Greece, Turkey and Italy. This rather highlights the fact that the online media in our sample seemed to focus on the perspective of quantity. The second most common L1 collocate in Greece, was ‘management’ (10% of the relevant articles), while in Germany it was ‘Syrians’, which appeared in 6.3% of the examined articles. In this regard, one could argue that the press in Greece focused on the managerial aspect of the refugees arriving in the country, while the selected online newspapers in Germany provided a geographical qualifier. In the UK, the second most common L1 collocate was the word ‘child’ (21.5%), a fact that underlines the emphasis given by the British press to the sensitive issue of child refugees. In addition, the L1 collocate ‘unaccompanied’ appeared before the word ‘refugees’ in 5% of the articles examined in the British press. The Guardian, reporting on the change of attitude of former Prime Minister David Cameron towards the ‘unaccompanied children’ from Syria, questioned whether ‘David Cameron’s U-turn on unaccompanied child refugees [should] be celebrated’ (Gentleman 2016). A different article in the same paper noted that ‘Britain will take thousands of unaccompanied child migrants from EU camps’, as a result of the recent vote in the House of Lords ‘to amend the immigration bill in order to require the government to let the children, currently in Europe, come to Britain’ (Gentleman and McVeigh 2016).

Unsurprisingly, during our analysis, we noticed that the word ‘hotspots’ was among those that most often appeared before the words ‘immigrants’, ‘refugees’ or ‘asylum seekers’ in the Greek newspapers Kathimerini and To vima (7.7%). The online coverage of both Greek newspapers condemned the ‘extremely poor living conditions’ in the hotspots, the ‘violent scenes’ among immigrants/refugees and the process of ‘immigrant/refugee registration’, while reporting/commenting on the accumulation of refugees in Greece, and ‘the ongoing tensions among the refugees’ within hotspots. To vima (2016a), presenting the conclusions of a Human Rights Watch report, stressed that ‘the refugee hotspots are unsafe and unsanitary’, while underpinning that ‘women and children in particular are left unprotected and Greek police appear to be indifferent to their plight’. Moreover, Kathimerini described the situation at the makeshift camp in Elliniko as a ‘health time bomb’, while underlining that not a single day passes without clashes (Georgiopoulou 2016). Of course, the ‘inhumane’ living conditions in the refugee camps were often reported by the German and British press as well.

In an article with the headline, ‘Refugee camps on Greek islands “bursting at the seams”’, The Telegraph wrote that ‘living conditions are rough, food is often inadequate and thousands of children, many of them Syrians who fled the war, are missing out on their schooling’ (Squires 2016), highlighting the lack of moral and safe conditions in the camps. In the same vein, Die Welt (2016a) reproduced pictures of the muddy landscape at the makeshift camp in Idomeni, where refugees’ tents were underwater. In its article, Die Welt argued that Idomeni was a place of drugs, prostitution and violence.

As listed in the table below, the third most used L1 collocate in the British press was the word ‘Syrian’. Significantly, immigrants were not so frequently described as ‘irregular’ or ‘illegal’ in our Greek and German samples (3.1% in Greece and 1.2% in Germany), something that does not apply to the British sample: 10% of the instances of the word ‘immigrant’ were preceded by the word ‘illegal’ and 6% by the word ‘irregular’. The latter collocate did not concern the words ‘refugees’ or ‘asylum seekers’, as by definition refugees and asylum seekers cannot be described as illegal/irregular. The German newspapers analysed also seemed to also focus on unaccompanied children who had been forced to abandon their countries (‘young’ and ‘minors—children’). Significantly, Die Welt (2016b) underlined that the cost of supporting the unaccompanied underage refugees is insurmountable for the German local authorities, invoking mention of the German Association of Towns and Municipalities. Furthermore, the same newspaper hosted articles which expressed concern about the living conditions of children from Syria at the hotspots in Greece, as well as the fact that they have been cut off from education. In another report, Die Welt (2016c) put forward the extremely odious issue of refugee child prostitution, describing it as a ‘catastrophic condition’ that the refugee children face. At the same time, various other terms appeared in the selected Greek newspapers just before our keywords, such as ‘refoulement’, ‘registration’, ‘relocation’ and ‘integration’ (‘to integrate’), that indicated discussion of the procedural actions relevant to migration, while on several occasions migrants, immigrants, refugees and asylum seekers were described as ‘trapped’ in the Greek territory.

Another important element that can be seen from Table 3 is the frequent use of the metaphors of ‘flows’ and ‘influx’ in the Greek, German and British press, highlighting the high volume of migrant and refugee arrivals. Although this seems to suggest a generally negative representation of migrants/refugees (Khosravinik 2009), this was not the case in our sample. At this point, before proceeding to the c-collocates (Table 4), it is worth noting some general assumptions regarding the examined online newspapers’ discourse on the refugee crisis. In this context, the sample of online newspapers tended to be favourable towards refugees. The latter, inter alia, were represented as victims of the civil war in Syria, and as helpless and desperate. Last but not least, the personal stories of the migrants and refugees were shared extensively by the online media analysed, adding a dramatic tone to the presentation of the news.

Table 4 C-collocates of ‘immigrants’, ‘refugees’ and ‘asylum seekers’ (top 30)
Full size table

C-collocates

The most common ‘c-collocate’ in the Greek newspapers Kathimerini and To vima was ‘Greece’, usually accompanied by the adjectives ‘Central’, ‘Northern’ or ‘Southern’ in order to provide a geographic location for the migrants’ and refugees’ arrival, and their accumulation, registration and relocation, as well as to present other incidents related to migrants/refugees (e.g. violent scenes, living conditions). Similarly, the German online media referred to ‘Greece’ to a significant extent according to Table 4 (25.8%). In general the selected German and British newspapers were sympathetic towards Greece, which seems to have been the EU country most affected by the refugee crisis, hosting tens of thousands refugees and irregular migrants, while a humanitarian crisis has been sparked there. However, the problems of accommodation and care for the migrants and refugees were issues that were also highlighted, and were touched upon by the Greek newspapers as well. Nevertheless, the most common c-collocate in Die Welt and Süddeutsche Zeitung was ‘Turkey’ (46.8%), something that interestingly applies to the British newspapers as well (69%).

It seems that all six online newspapers paid a lot of attention to the EU–Turkey deal, referring to the nature and provisions of this agreement, as well as to the perspective of its implementation. For instance, Süddeutsche Zeitung (2016) expressed ‘hoffnung und Skepsis nach Flüchtlingspakt mit der Türkei’ [hope and scepticism about the refugee pact with Turkey]. What is worth noting is that the word ‘controversial’ was very often deployed to describe this accord. Various articles included politicians’ statements stressing the importance of the implementation of this deal, while international organisations such as the UN High Commissioner for Refugees and Amnesty International raised their voices against the very substance of the agreement, expressing their scepticism regarding the legality of the planned deportations. At the same time, they did not neglect to underline that Ankara must respect the relevant provisions found in international law regarding the protection of refugees. Last but not least, serious concerns about a possible collapse of the EU–Turkey agreement were unanimously expressed in the newspapers studied, while the Turkish threats regarding its implementation were reflected as well.

Turning our attention to the coverage of this crisis by the Greek newspapers in our sample, they seem to focus on the state of play at the hotspots within Greece: at the makeshift camp in Idomeni where violent incidents and protests have taken place, and at the port of Piraeus where several migrants and refugees have been trapped in makeshift camps. ‘Hunger strike in Chios’ where ‘the situation has been derailed’, ‘turmoil in Idomeni’, ‘the deadlock in Idomeni remains’, and ‘stone-throwing in overcrowded Idomeni’ were some of the headlines in Kathimerini (2016a, b, c). On the other hand, To vima (2016c) likened the situation in Piraeus and Idomeni with the living conditions in favelas. Of course, the transfer of refugees from the improvised camp at the port of Piraeus was an issue of high concern, which was covered by all the newspapers examined. For instance, The Guardian wrote that ‘Athens is under pressure: city races to clear port’s refugee camp before tourists arrive’ (Smith 2016).

Furthermore, both in Greece and in Germany, the word ‘borders’ appeared frequently, while the British press often referred to the ‘Greek–Macedonian border’. The word ‘FYROM’Footnote 2 appeared in 11.9% of the articles analysed in the Greek papers. These references to ‘FYROM’ were mostly made in the context of the closing of the Balkan route or the violent scenes between the authorities of the Former Republic of Macedonia and the refugees and migrants at the borders between Greece and that country. The British press also highlighted that clashes took place. The Guardian reported on several injuries after ‘clashes at Greece–Macedonian borders’, when ‘Macedonian police fired teargas and rubber bullets at crowds on the Greek side of the border’ (The Guardian 2016). On 11 April 2016 The Telegraph presented the strong opposition of Alexis Tsipras to the above-mentioned practices of the authorities from the Former Yugoslav Republic of Macedonia, writing that ‘Greece’s prime minister accused Macedonia of “shaming” Europe after the Balkan country’s police used plastic bullets, stun grenades and tear gas to beat back refugees from a border fence’ (The Telegraph 2016). At the same time, the critical reaction of President of the Hellenic Republic Prokopis Pavlopoulos concerning the stance of the Former Yugoslav Republic of Macedonia in Idomeni was noted in both Kathimerini and To vima. ‘FYROM has no place in the EU and NATO’ (To vima 2016b) said Mr Pavlopoulos. In general, both Kathimerini and To vima published articles or statements which expressed their scepticism and disapproval regarding the attitude of the authorities of the former Yugoslav Republic of Macedonia at the border between it and Greece.

Importantly, all the newspapers in the study made mention of the Greek islands, and more specifically of Chios and Lesvos, where the numbers of migrant and refugee arrivals rocketed. Additionally, from an examination of Table 4, it seems that there were several references to Italy in the Greek and the British press, something that cannot be said for the German press in our sample.

At this point, it is interesting to note that when analysing the articles from the two British newspapers, the authors of this article expected a significant reference to the Brexit debate on migration. However, references to the word ‘Brexit’ or to the EU referendum appeared in only 4% of the articles examined.

Finally, we were surprised to note that the articles studied in the Greek and the German press did not include in a percentage worth mentioning the words ‘terrorists’, ‘terrorism’, ‘ISIL’, ‘IS’ or ‘ISIS’.Footnote 3 This contrasts with the British press, where the above-mentioned terms emerged in 17% of the articles examined. The authors of this article can only assume that the almost complete absence of any reference to terrorism in the German and Greek media is part of the wider effort of (opinion) leaders to tackle the spike in hate crimes in the two countries.

Conclusions

The findings of this article suggest that with Europe shaken to its foundations and the EU’s values and common future having been called into question, the media seems to have taken a uniform approach to covering this crisis. This contrasts with the conflicting views of Europe’s leaders. At the same time, it seems that hate speech has caught the attention of the media and caused public concern. As mentioned earlier, refugees were portrayed as victims of the civil war in Syria, helpless and desperate, while there was also a focus among all the newspapers on the ‘numbers’ of the crisis. Unsurprisingly, the EU–Turkey agreement was at the heart of the daily coverage. Apart from references to the nature and provisions of this agreement, as well as its implementation, various articles expressed scepticism regarding its legality and concerns about its possible collapse. Finally, all the newspapers examined paid a good deal of attention to the tragic issue of child refugees, who are living in extreme peril and facing inhumane living conditions and various atrocious threats.