Refugees in the News: Comparing Belgian and Swedish Newspaper Coverage of the European Refugee Situation during Summer 2015

Aim and theoretical framework This comparative content analysis of Belgian and Swedish newspaper coverage of the ‘refugee situation’ in 2015 revolves around responsibility indicators, news actor characteristics, and thematic emphasis. As a potential influential factor in the public opinion formation process, studying media portrayals is an essential first step. The two countries differ in terms of real world indicators. Sweden scores far better on the Migrant Integration Policy Index (MIPEX) and the actual number of incoming refugees is higher. Belgium and Sweden also differ when looking at general and journalistic cultural values. Therefore, they form an excellent case study to examine intercultural differences and resemblances in reporting on refugees. In her work, Szcepanik (2016) illuminates the existence of a normative ‘refugee archetype’ and imagined refugeehoods in media coverage by distinguishing between ‘good, genuine’ refugees and ‘bad’ refugees. A ‘good refugee’ is female, poor, helpless and has a particular nationality. Prominence in media texts is an indicator of who deserves protection and who does not, a classification in direct conflict with the definition of a refugee as stated in the Geneva Convention. ‘Bad refugees’ are depicted as people who intend to abuse the social welfare system in welcoming countries in Europe and lack the attributes of good refugees. Previous research by Van Gorp (2005) showed that refugees were most often portrayed as victims or intruders in the Belgian press. The language of blame attribution and the suggested solutions to counter problems for which blame is insinuated is an important focal point in news research. In British newspapers, for instance, 17% of the articles on refugees mentioned the words terrorists, terrorism or a reference to IS. These references were nearly non existing in the Greek and German press (Fotopoulos & Kaimaklioti, 2016). Methodology We conducted a multilingual quantitative content analysis in Swedish, Flemish and French-language newspapers in Belgium. As the number of first-time asylum seekers in the EU peaked during the third quarter of the year (Eurostat, 2016b), we coded all news articles on refugees appearing from August 24 until September 27, 2015 (N= 898). This period contains events as Merkel’s ‘Wir schaffen das’, the death of Alan Kurdi, reports on the dangerous journey towards Europe and countries closing their borders. For each media landscape, the main national quality newspaper and the largest popular newspaper were analyzed: De Standaard and Het Laatste Nieuws (Flanders); Le Soir and L’Avenir (French-speaking readership in Brussels and Wallonia), Dagens Nyheter and Aftonbladet (Sweden). A coding instrument was developed and adapted during multiple rounds and pre-tested. Intercoder reliability was measured on a sample of Swedish and Belgian news articles. Research hypotheses H1: Immigration figures are more often mentioned and discussed in the Belgian than in the Swedish press. H2: Swedish newspapers a. report more on welcoming reactions in civil society and politics, rights, humanitarian dimensions, success stories of refugees, and post-arrival integration. b. use fewer threat-related news items than Belgian newspapers do. H3: Swedish newspapers report more on female refugees. H4: a. Refugees are less often quoted and paraphrased than other actors. b. Refugees are more often quoted and paraphrased in the Swedish press. H5: a. Refugees are primarily presented as a collectivity instead of as individuals. b. Refugees are less presented in a collective way in the Swedish press. Results Immigration figures (H1) The Belgian corpora contain more references to immigration numbers than the Swedish corpus (38.1% versus 27.3%, X²=10.611, p=0.001), but this difference is more outspoken because of the higher frequency of references in the southern part of Belgium (50.6% in Wallonia versus 28.1% in Flanders), (X²=30.367, p=0.000 < 0.001) Positive/negative reactions (H2) Hypothesis 2 is only partially true: H2b can be confirmed, but H2a is to be rejected if we take the Flemish and French-language corpora together. If separated, we note how the Swedish media do report more positive reactions to the refugee issue than the Flemish newspapers, but not more than the French-language texts. Gender (H3) Of the 299 individual refugees mentioned, 30.4% and 17.4% are female in the Swedish and Belgian press respectively (X²=6.295, p=0.009). No difference between the Flemish-speaking and French-speaking press could be found. Quotes (H4) Refugees are less often quoted than non-refugees (0.32 versus 1.60 times per article, t=-0.177, p=0.000). Refugees are less often given a voice than non-refugees, but they are more often directly quoted than paraphrased. The phenomenon is more outspoken in Belgium than in Sweden and especially outspoken in Flanders. Collectivization (H5) The rate of collectivization (1=very collective; 5=very individual) is not different between Belgium and Sweden (1.51 versus 1.46, t=0.708, p=0.479) and within Belgium (1.46 Flanders versus 1.58 in the French-language press, t=-0.1369, p=0.172). Conclusion and discussion Our analysis made it clear that the Belgian news coverage shows regional diversity, as the Francophone Belgian journalists show more tolerance towards migration and tend to lean more towards their Swedish counterparts, both focusing more on positive themes than their Flemish colleagues. What is apparent, however, is the lack of voice given to the refugee actor in all news texts examined. Other actors, mostly elite politicians, were quoted and paraphrased much more frequently than the protagonists themselves. This lack of voice appears to be consistent and merits the attention and action of the media professionals.status: accepte