Master’s Thesis
Graduate School of Communication
Master’s programme: Erasmus Mundus Journalism
Refugees in the News:
An Analysis of Frames in German Regional Newspapers Tobias Widmann
University of Amsterdam
Student number: 10831746
Supervisor: Dr. Sophie K. Lecheler Friday, 26th June 2015
Abstract
The issue of asylum seekers is an incendiary topic around the globe, in which the media play an important role, since media has the power to influence public opinion on refugees.
Germany, like many other European countries, is facing growing anti-immigration sentiments among the population. However, we do not know which role the media play in these
developments and whether these sentiments have effect on the news framing. The aim of this study is to analyze the framing of refugees and asylum seekers in German regional
newspapers. This study applies a deductive framing analysis in a longitudinal research design with frames previously identified in the news coverage of other European countries. Findings suggest that frames on refugees tend to be more positive, and that negative valenced frames found in other European countries might not be applicable to German news coverage. German newspapers do not frame newcomers in relation to Islam and do not portray the ‘other’ culture as a threat. Furthermore, the results show that journalists tend to portray refugee more positively when they write about individuals. Lastly, there is little systematic evidence that the anti-immigration movement Pegida influenced the framing and that coverage differs between East and West.
Refugees in the News:
An Analysis of Frames in German Regional Newspapers
The rising numbers of refugees and asylum seekers entering the European Union represent an important challenge, not just for the municipalities who need to provide shelter and services for the new arrivals, but also for the society as a whole. Many Europeans are finding themselves suddenly confronted with seeming 'strangers', which can arouse suspiciousness, prejudices and conflict on the one side, but also helpfulness, respect and cooperation on the other. Given the importance of the refugee issue nowadays in Europe, this study focuses on the examination of the framing of refugees in German regional newspapers. This is important since the way the media frame this important topic has the power to
influence public opinion and to create support for the formation of popular movements (e.g., Freeman, Hansen, & Leal, 2013; King & Wood, 2001).
Therefore, this study tries out several different frames, which have been previously identified in the news coverage of other European countries. The quantitative content analysis with a longitudinal design analyzes whether these frames are also applicable to German coverage and investigates possible effects of the Pegida movement on framing. Only a few studies, if any, have applied the concept of framing of refugees in the German case.
Uniquely, this study focuses on regional newspapers, an area rarely ever analyzed in relation to framing. Previous research on immigration and especially irregular immigration shows a tendency of focusing on national news outlets, to facilitate cross-national comparison
(Thorbjørnsrud, 2015; Benson, 2010; Vliegenthart & Roggeband, 2007). However, especially immigration is an area where the local perspective is relevant, because immigrants are part of the daily life in local communities. Regional newspapers cover the daily life of citizens closer than national newspapers, which emphasize broader considerations and implications (Lawlor, 2015). This study therefore tries to contribute to the knowledge about framing refugees by
delivering results from an East-West comparison of the German regional press, which has an impressive readership of more than 30 million readers a day (Pasquay, 2014).
Studying Frames in the News
This study focuses on frames in news media. However, the increasing popularity of the framing concept in mass communication research has gone hand in hand with an increased ambiguity in the application of the framing concept (de Vreese, 2005). Entman (1993) referred to framing as “scattered conceptualization” (p. 51) since studies often lack a clear definition and use a context-specific operationalization instead of a generally applicable one (Scheufele, 1999). Kinder (2007) states, that even though the term ‘frame’ is been used in communication science for several decades now, it is still not clear what actually
constitutes a frame. Therefore, the following part of my paper presents a conceptual definition of frames and a critical discussion of the theoretical assumptions, on which this paper is based on.
Framing Theory
The concept of news framing derives from two fields, namely psychology and
sociology. In psychology, the “Asian decease”-studies by Kahneman and Tversky (1984) are often seen as the basis to contemporary framing theory. In their studies, Kahnemann and Tversky gave important insight into the psychological processes of framing, by showing decisions of people can be changed through the presentation of information in semantically different ways (Lecheler, 2010). Their results present the base for the so-called prospect theory, which states that people tend to be risk seeking when expecting losses and risk averse when expecting gains (de Vreese, 2002).
In sociology, the term has already been used since the 1950's (Bateson, 1955). Bateson argues that statements do not possess an inherent meaning, but that they do obtain their meaning through a frame, which is constituted through context and style (Vliegenthart
& Van Zoonen, 2011). Goffman (1974) argues along this line by saying that meaning is only generated through interaction, contextualization and interpretation, which lead to “social frameworks” (p.24). Social frameworks define which actors, issues or events are relevant or appropriate for us and which are not (Vliegenthart & Van Zoonen, 2011).
The adaption of framing in the field of communication science began mostly in the early 1990's with the publication of Entman's (1993) article 'Framing as a Fractured
Paradigm', and both the sociological and psychological approaches are now being absorbed into the news framing theory. Since most citizens rely on mass media for information about politics, news frames play the most important role in forming and altering opinions and political attitudes (de Vreese, 2002).
By offering ways of classifying information and increasing salience for certain aspects, frames affect individuals' pattern of thoughts. Iyengar (1987) describes framing effects as “changes in judgement engendered by subtle alterations in the definition of judgement or choice of problems” (p. 816). In political communication, studies of framing effects often focus on how news can affect our thinking of an issue (Scheufele, 1999). Different recent studies hint to the effects news frames have on the audience and on public opinion. Druckman, Peterson and Slothuus (2013) for instance, point out that the effects of polarization in party politics on the formation of public opinion. Boukes and colleagues (2014) showed the effect of human interest framing on political attitudes and the way people attribute responsibility of an issue.
Regarding framing effects in relation to minorities, studies have shown that frames about immigration can have detrimental effects on attitudes and opinions towards immigrants (Van Londen, Coenders, & Scheepers, 2010; Boomgaarden & Vliegenthart, 2009).
Consequently, content analyzing the predominant frames can be considered a “prerequisite for the study of effects of news frames” (de Vreese, Peter, & Semetko, 2010, p. 108). It is
relevant to understand how media frame refugees to understand the effects these frames might have on public opinion towards refugees and asylum seekers.
What is a ‘News Frame’?
Taking a broader sociological approach on frames, frames are often described as “a central organizing idea or story line that provides meaning to an unfolding strip of events, weaving a connection among them” (Gamson & Modigliani, 1987, p. 143). Tuchman (1978) says that a frame “organizes everyday reality” (p. 193) and by doing so, providing the parameters “in which citizens discuss public events” (p. IV). Frames therefore serve as “conceptual tools which media and individuals rely on to convey, interpret and evaluate information” (Neuman, Just, & Crigler, 1992, p. 60). Entman (1993, p. 52) on the other hand emphasizes the selection aspect of frames and argues that to “frame is to select some aspects of a perceived reality and make them more salient in a communicating context, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described”.
Based on the definitions above, frames can be described on the one hand as a tool for the audience to classify and interpret information efficiently (Lecheler, 2010). On the other hand, news frames “stress certain aspects of reality and push others into the background: they have a selective function. In this way, certain attributes, judgments, and decisions are
suggested” (Lecheler, 2010, p. 4).
In political communication, frames can be found in different locations and different processes. Kinder and Sanders (1996) for example distinguish between 'frames in political discourse' and 'frames as structures of the mind', which are described as “cognitive structures that help individual citizens make sense of the issues that animate political life” (p.164). De Vreese (2002) portrays a similar distinction between 'media frames' and 'audience frames'.
Scheufele (1999) also differentiates between 'media frames' and 'individual frames', since frames are schemes for both, “presenting and comprehending news” (p. 106).
However, is this distinction between media frames and individual frames meaningful? Regarding previous research, a majority of framing studies focused either on media frames or on possible effects of media frames (e.g. Semetko & Valkenburg, 2000; de Vreese et al., 2010; Boukes et al., 2014; Lecheler, Keer, Schuck, & Hänggli, 2015). Van Gorp (2005) for instance, argues that there is no such thing as an 'individual frame', because frames are not located independently in individuals but in our common culture. Here, Van Gorp follows the line of Goffman (1981), who argued that frames are “a central part of our culture and are institutionalized in various ways” (p. 63). Van Gorp (2005) refers to these patterns of thoughts not as 'individual frames', but instead as 'schemata' (p. 487). Scheufele and Iyengar (2012) agree by saying that “schemas are culturally shared” (p. 15). As a result, tests of framing effects within a cross-cultural comparison will never achieve as strong applicability effects as in one given culture (Scheufele & Iyengar, 2012). Hence, readers as well as journalists are constantly accessing frames that are present in the repertoire of our common culture. This study also focuses on media frames present in texts (words, phrases, etc.), and follows the line of Van Gorp by arguing that these frames cannot exist independently of the culture they are embedded in. Therefore, this study is testing whether frames, which have been identified in the cultural context of different countries, are applicable in Germany. Measuring Frames in the News
The chapter above provided a conceptual definition of news frames. However, to identify news frames, an operational definition is needed as well. Regarding the identification of frames, there is unfortunately also little consensus on the method (Matthes, 2009). One approach is inductive in nature. Inductive studies are more qualitative and analyze news stories with an open view in order to identify possible news frames, with only roughly
defined preconceptions of the frames (Semetko & Valkenburg, 2000). Thus, frames emerge during the analysis of the material. Recent studies with an inductive approach detected for instance frames in the coverage on homicide of women (Richards, Gillespie, & Smith, 2013) or in relation to female genital cutting (Sobel, 2015). However, this method is often criticized for being too labor intensive, relying on a small sample and being too difficult to replicate (Hertog & McLeod, 2001).
The second approach is a deductive content analysis, which analyses the presence of frames, which are defined and operationalized prior to the investigation (de Vreese, 2005). Scholars have argued in favor of the second approach, using a priori defined
operationalization of frames in content analysis (de Vreese, 2005; Capella & Jamieson, 1997). This approach can be replicated easily, can analyze a large sample and detect differences in framing between and within media (Semetko & Valkenburg, 2000). Recent deductive studies analyzed for example the coverage on environmental risks in Chinese newspapers (Tong, 2014), or the coverage of the 2009 European parliament elections (Schuck, Vliegenthart, & de Vreese, 2014). This study also applies a deductive approach, since, in order to assess differences in framing over time and between media, a large sample is necessary.
There are multiple ways in which frames can be applied in news articles. However, scholars agree that there are certain characteristics, which can be captured in the deductive analysis of news articles (de Vreese, 2005). Iyengar (1991) for example distinguishes between thematic framing, which refers to more analytical and contextual news reporting, and episodic framing, which refers to specific events or individuals. However, Vliegenthart and Van Zoonen (2011) criticize the level of confusion in the application of different frames and the use of different labels for similar frames. They argue for example, that there is
considerable overlap, for instance, between Iyengar’s (1991) ‘episodic’ framing, and de Vreese’s (2005) ‘human interest’ frame.
Issue-specific frames are a specific group of frames, which only pertain to specific topics or events. If issue-specific frames contain positive or negative elements, they are called 'valence frames'. Valence frames are considered as “indicative of “good and bad” and
(implicitly) carry positive and/or negative elements” (de Vreese & Boomgaarden, 2003, p. 363). Examples of issue-specific media frames with inherent valence from other topics are the 'anti-feminism' frame (Terkildsen & Schnell, 1997), or the 'risk' and 'opportunity' frames (de Vreese, Boomgaarden, & Semetko, 2011).
On the other hand, generic frames transcend thematic limitations and are broadly applicable to different news topics, times and cultural contexts (de Vreese, 2005). Semetko and Valkenburg (2000) for example, argue that news are mostly framed in five different
generic frames: 1) conflict frame, 2) human interest frame, 3) economic consequences frame,
4) morality frame and 5) responsibility frame.
In regard of generic frames, Van Gorp (2007) argues, that they have been analyzed in such a general way that they are better referred to as schemata or genres. The analysis of issue-specific frames is similarly problematic because of the limited opportunities they offer for generalization and comparison (De Vreese, 2005). Acknowledging the shortcomings of both, generic and issue-specific frames, this study applies a wide range of issue-specific frames concerning refugees and asylum seekers frames from different levels of society, e.g. economy, politics and culture, which will be discussed in the following chapter. And in addition, this study also analyzes the usage of two generic frames, episodic and thematic, to research whether the coverage on refugees is focused on individuals or on the broader
picture. By applying both issue-specific and generic frames, this study will be able to provide both an in-depth as well as a style-related insight into refugee coverage in German
newspapers.
As de Vreese (2012) points out, the challenge in framing research is to translate and operationalize these general definitions in empirical studies. The fundamental question remains: what (which components) constitutes a frame? The big aim and necessity therefore is, to make it absolutely transparent, how the different components are aggregated to
represent frames (de Vreese, 2012). Thus, each study must be precise on the components and items constituting a frame, conceptually and operationally (Matthes, 2009). Hence, the following chapters present a conceptual as well as an operational definition of the frames included in this analysis.
Minorities in the Media
Topics related to ethnic minorities, immigration and immigrants have become more important in various fields of the society from the early 1990s. In European politics these issues have shifted from ‘low politics’ to ‘high politics’, and the volume of media coverage of the topic has increased in European countries (Ter Wal, 2002). This is important, because exposure to news can explain a possible rise in anti-immigration attitudes (e.g., Oliver 1999, Vergeer, Lubbers, & Scheepers, 2000). For a majority of citizens, media offers the only possibility to define their perception of minority groups (Spoonley & Trlin, 2004). Bobo (1997) argues that “images of racial minority groups in the media at once reflect and shape the attitudes and beliefs others will hold toward those groups” and that therefore “it is important to assess whether the general media environment is one that discourages thinking of racial inequality” (p. 7,8).
Several studies dealt with the framing of immigrants in the media. A study by Van Dijk (1983) from the Netherlands shows that coverage in media is mostly a reflection of viewpoints of the majority in society and that minorities are given only little opportunities to express their opinion. News coverage about minority groups mostly conveys a negative
image through a narrative of 'illegality', related to sensational conflicts or crime coverage (Ruhrmann, 2002; Gabrielatos & Baker, 2008; Horsti, 2007; Sulaiman-Hill, Thompson, Afsar, & Hodliffe, 2011; Suro, 2011). A study from Germany (Gräf, 2009) presents similar results, showing that the most often mentioned topics in connection to minorities are crime and terrorism. The reflection of difference and promotion of 'otherness' is potentially more marketable, since conflict is often considered more newsworthy (Shoemaker, 2006).
Dominant negative discourses can perpetuate social distance, portraying ethnic minorities as threats and framing interaction between minority and majority groups as conflicted (Haynes, Devereux, & Breen, 2004).
Framing Refugees and Asylum Seekers
As the above mentioned studies show, there is a multitude of research done on the framing of minorities in the media. However, specific studies on the framing of asylum seekers in Europe as a subgroup of immigrants are not as numerous as one might assume given the importance of this issue (Horsti, 2007). What is more, studies that deal with the framing of refugees and asylum seekers in German newspapers are rare.
According to the 1951 Refugee Convention, a refugee is someone who "owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality, and is unable to, or owing to such fear, is unwilling to avail himself of the protection of that country." The terms asylum seeker and refugee are often confused: an asylum seeker is someone who says he or she is a refugee, but whose claim has not yet been definitively evaluated. According to the United Nations Refugee Agency (UNHCR), about 1 million people seek asylum on an individual basis every year. In mid-2014, there were more than 1.2 million asylum seekers worldwide (UNHCR, 2015).
Ever since the beginning of the Syrian Civil War in early 2011, the numbers of asylum seekers in Europe multiplied. Together with refugees from other conflict areas like Afghanistan, Iraq or Eritrea, the EU faced a sharp increase of asylum applicants. In 2010, 240,410 people applied for asylum in the 27 member states of the EU. This number rose to almost 450,000 asylum applicants in 2013 in the EU-28 states (Ec.europa.eu, 2015). In Europe, Germany receives the largest absolute number of asylum applications. In 2010, 41,332 new asylum applications were counted in Germany. In 2013 the number rose to 109,580 new applications and reached a peak in 2014 with 173,072 applications (Bamf.de, 2015a).
The rise of asylum seekers brings up questions about the integration of refugees in Germany and in other European countries. Different anti-immigration movements came to existence all over Europe and right-wing populist parties receive high voter support. Populism is on the rise and at the same time, the support for more radical anti-immigration parties is increasingly prominent in many European countries (van Spanje, 2010).
In Germany, the so-called Pegida-movement ('Patriotic Europeans against the Islamisation of the West') was founded on the 20th October 2014 in the city of Dresden in Eastern Germany and since then received a vast quantity of media coverage, nationally as well as internationally. The movement denies being radical or racist and says it fights against what it sees as the growing influence of Islam in Europe. Since the 20th October 2014, the organization is holding regular Monday marches in Dresden, demonstrating for stricter enforcement of asylum laws. A survey conducted by the German public broadcaster ARD showed that attacks on refugees or refugee centers multiplied since the beginning of Pegida (swr.online, 2015).
A majority of studies dealing with the framing of refugees contained two groups of frames: positive and negative. As mentioned above, issue-specific frames, which contain
notions of positive or negative evaluations, are called 'valence frames'. In relation to the coverage on refugees, valance frames mostly divide refugees into genuine, political refugees on the one side and bogus, economic refugees on the other side (Robinson, 1996; Kaye, 2001; King & Wood, 2001; Crawley, 2005; Finney, 2005; Nickels, 2007; Every &
Augoustinos, 2008; Steimel, 2010). Issue-specific frames on refugees and asylum seekers fit into either a positive/legitimate or negative/illegitimate group (Nickels, 2007). However, unfortunately, available framing studies on the subject have chosen to translate these two groups into an array of frames with different names. To establish, which frames are relevant for the German case, this study examines eight different frames established previously. Six of these frames are relevant, because they have been specifically established in reference to immigration and refugee news coverage. This enables an in-depth analysis of the content of news article on the topic. A second group of two frames are broader or generic in nature, focusing more on the general narrative style of the article. They are important since the narrative of the article can have implications for citizens' attribution of responsibility, their policy views (Iyengar, 1991) and their emotional reactions towards an issue (Gross, 2008).
The six issue-specific frames are the most common frames related to refugee portrayal in the literature (Van Gorp, 2005; Nickels, 2007; Roggeband and Vliegenthart, 2007; Ettinger & Udris, 2009). These frames, illustrated in table 1, constitute three pairs of two frames each, which counter each other in their valence. The frames are bogus-refugee
frame vs. humanitarian frame, the intruder frame vs. the victim frame, and the Islam-as-a-threat frame vs. the multicultural frame.
< TABLE 1 ABOUT HERE >
The bogus-refugee frame and the humanitarian frame are based on the distinction between ‘genuine’ and 'bogus' refugees (e.g. Ettinger & Udris, 2009, Nickels, 2007). According to Van Dijk (1997), the concept of 'economic refugees' proved to be a powerful
rhetorical strategy in politics to establish stricter refugee policies and immigration controls. The media differentiates similarly between genuine and bogus refugees and asylum seekers (Nickels, 2007). For instance, the British media portrayed East Germans fleeing to West Germany during the Cold War as political refugees (McLaughlin, 1999). After the fall of the Berlin Wall, the phrasing changed, leading the British media to portray people who fled from the East as being part of a larger migration problem rather than as political refugees. Coole (2002) argues that this change in political rhetoric enabled “a climate of unease and hostility towards people seeking asylum” (p. 839), and media began to focus on the perception that the majority of refugees are not in genuine need of protection from persecution. The news began to portray refugees as people receiving social benefits, taking away jobs and are not
deserving state protection (Coole, 2002). This bogus-refugee frame is however often counter-framed by the humanitarian frame (Ettinger & Udris, 2009). The humanitarian frame depicts refugees clearly as victims of wars and thus as 'genuine' refugees, who need immediate protection.
The so-called victim frame by Van Gorp (2005) was identified in the Belgian press. In this frame, a refugee is considered as a vulnerable victim “who is outside his/her country of nationality or habitual residence because of a well-founded fear of persecution because of his/her race, religion, nationality, membership in a particular social group or political
opinion” (Van Gorp, 2005, p. 489). The refugees are portrayed as incapable of changing their circumstances, as being forced to leave their home, as poor and as unknowledgeable about the system and language in the western world. Conversely, the counter-frame, the intruder frame depicts asylum seekers as suspects or even as the associates of human traffickers, as criminals who abuse the asylum system and as an endangerment for the locals (Van Gorp, 2002).
Vliegenthart and Roggeband (2007) identified two frames in the Dutch press, which depict refugees in opposite ways. The Islam-as-a-threat frame portrays the culture and religion of refugees as a problem, especially the Muslim religion. In this definition, refugees are portrayed as a threat to the values of our western societies and connections are being made between refugees and terrorism. The increase of this frame was fostered by the events of 11th September 2001. The counter-part of this frame is the multicultural frame, which portrays cultural diversity as an asset that leads to benefits to the western societies. This frames emphasizes the dialogue and respect between locals and refugees from different cultures (Vliegenthart & Roggeband, 2007).
Distinction between Positive and Negative
This study takes a similar view that frames about refugees are either assigned to a 'positive' or 'negative' group. The positive group consists of frames, which convey a more liberal standpoint towards refugees and different cultures, meaning that asylum seekers are framed as genuine, victims of war who flee war or another humanitarian catastrophe. The negative group focuses on the danger of immigration, the abuse of the asylum system and possible threats through refugees or their culture and religion.
Previous literature does not provide knowledge about which frames cover the topic the best. Frames have been identified and developed in different cultural contexts and media systems. Therefore, this study applies a multitude of frames to investigate their applicability in the German news coverage.
Some scholars argue, that there is a predominance of negative frames in the coverage on refugees. Figenschou and Thorbjørnsrud (2015) for example claim that much of the
literature points to a predominance of an 'illegality frame' in the framing of refugees, in which irregular immigrants are framed as bogus refugees, as threats or as lawbreakers. This aligns with the results of, for instance, Benson (2013), who found that immigration media debates in
France and the US emphasize the law-and-order frame, which focuses on national security and a need for stricter law enforcement in regards to refugees. Kim and colleagues (2011) also found a predominant enforcement frame in their study from the US.
On the other hand however, Figenschou and Thorbjørnsrud (2015) suggest that the mainstream news media have a more liberal approach towards refugees than scholars suggest. Some scholars even hold the opinion that media stay silent about problems related to
immigration, thereby taking a pro-refugee standpoint (Thorbørnsrud, 2015). Several studies hint to a predominance of 'positive' frames related to refugees and asylum seekers. A study by d'Haenens and de Lange (2001) from the Netherlands for instance, concludes that the
portrayal of refugees in Dutch regional newspapers mostly tries to anticipate the feeling of the readers, by giving the stories a 'human face'. They showed that the 'human interest frame' was the most widely used frame in the coverage of refugees in Dutch regional newspapers. Van Gorp (2005) analyzed the usage of the victim and intruder frames in the Belgian press, and his results point out that the victim frame is predominant. Similar to the Belgian study, Steimel (2010) argues that refugees in the United States are either portrayed as victims or frauds. She concludes that the innocent victim story clearly dominates as the preferred narrative for refugees settling in the US.
The contradictory results regarding the predominant frames in the refugee and asylum seekers debate in Europe and the United States brings forward the question, how German media frame refugees. Based on studies analyzing the portrayal of immigrants and minorities in Germany (Ruhrmann, 2002; Ruhrmann & Sommer, 2005; Gardner, Karakasoglus, & Luchtenberg, 2008) one can assume that the framing of refugees in Germany tends to be more negative. However, Bauder (2007), for instance, argues that the German immigration discourse is more and more framed in a liberal and humanitarian way. Further, there is reason to believe that the coverage in Germany might be more positive than in other European
countries. In the Netherlands, for example, political actors like Geert Wilders or the murders of Theo van Gogh shaped a negative image of Islam and other cultures (Roggeband & Vliegenthart, 2007). However, right-wing populists or political murders have not been present in the political discourse in Germany. Therefore, one can assume that the media framing of refugees might tend to be more positive. This study tries to answer this question and to enlarge the previous research by identifying the dominant frames in the German regional newspaper coverage and to analyze whether positive or negative frames are dominating:
RQ1: To what extent are positive and negative frames used to cover refugees and asylum seekers in German regional newspapers?
Generic Frames as an Overarching Framework
As mentioned above, next to the six issue-specific frames, this study also applies on a second level and as an overarching framework the broader and generic episodic and thematic framing. Thereby, I investigate whether the portrayal of refugees is more focused on
individual persons or events or on a more statistical, analytical and contextual coverage. Iyengar (1991) for instance, argued that episodic framing leads to the audience blaming individuals for their problems. “Thus, the poor woman on welfare is held responsible for her fate, not the government or the system” (Semetko & Valkenburg, 2005, p. 96). In opposite to Iyengar's findings, several studies show a link between episodic framing (or human interest framing) and positive frames on refugees, which leads to positive effects on immigration attitudes. A study by Steimel (2010), concerning the portrayal of refugees in American human interest stories, shows that episodic framing does not necessarily lead to placing responsibility to the individual refugee, but instead to the system and to institutional factors by depicting refugees as innocent victims. Furthermore, Piazza and Haarman (2011) argue that episodic framing may lead to a positive picture of refugees and to a heightened sense of
community and solidarity. D'Haenens and de Lange (2001) argue that citizens exposed to episodic framing will have more positive sentiments towards the establishment of asylum seeker centers. The study by Cho and Gower (2006) shows that human interest framing increases empathy between readers and the involved parties. Further studies have
documented the so-called 'person-positivity bias' (Sears, 1983), which states that the public views generally individuals more positive than groups (Nilsson & Ekehammar, 1987; Iyengar et al., 2013).
Based on the literature above, this study tries to further understand the relation between thematic and episodic framing and positive or negative coverage on refugees. The question whether episodic framing is more positive or negative is important, since episodic framing causes higher emotional reactions (Aarøe, 2011; Gross, 2008). A study by Lecheler and colleagues (2013) about the role of emotions in framing effects on opinions about immigrations showed that emotional patterns correspond to the content of the frame. This means that positive frames elicit positive emotions. The study also showed that an increase of positive emotion might lead to more favorable opinions on immigration. Therefore, whether episodic framing is more positive or negative might affect the influence of framing of refugees has on immigration attitudes.
Taking the above mentioned literature on the linkage between episodic framing and positive evaluations of refugees and other minority groups as a point of departure, I suggest that the positive group of frames in this study is linked to episodic, while conversely, thematic framing is connected to a higher extent to the negative frame group:
Hypothesis 1: Episodic framing is positively associated with positive frames and thematic framing is positively associated with negative frames.
A further aim of this study is to investigate possible effects of the Pegida movement on the framing of refugees and asylum seekers. A study (swr.online, 2015) conducted by the ARD public-broadcasting news magazine ‘Report Mainz’ revealed that, since beginning of the Pegida marches, violence against refugees, migrants and refugee homes has increased by 130% in Germany. Whether or not these developments are reflected in the media portrayal of refugees, can have influence on the public's attitudes towards refugees. Koopmans (1996) has previously revealed a connection between right-wing violence and media coverage on the asylum debate in Germany. The results show that higher visibility of the asylum debate led to more extreme-right violence. Furthermore, Koopmans and Olzak (2004) showed that the visibility of anti-immigration movement activities in the media has a significant effect on the level of violence against refugees and ethnic minority groups. These results lead to the important question whether Pegida has influenced not only the visibility, but also the framing of refugees, since this could lead to a further increase of violence towards minorities.
In light with studies in other countries, I expect a rise of the level of negative framing within the coverage. Regarding the current situation in Europe, it is assumed that populism is on the rise, manifested in political parties as well as in anti-immigration movements, like Pegida. Kriesi and colleagues (2008) argues that the rise of populism can be explained by increasing anti-immigration sentiment among significant parts of the electorate. Different scholars argue that the media responds to this development firstly with an increased coverage on immigration issues (Boomgaarden, 2007; Schafraad, Wester, & Scheepers, 2009), and secondly with a more populistic debate (Schumacher & Rooduijn, 2013). Increased visibility of immigration topics in the media however, as mentioned above, can lead to a significant increase of support for anti-immigration parties (Karapin, 2002; Boomgaarden &
Vliegenthart, 2007). This again might have influence on the news coverage, which then again might lead to influence on the support for populist political parties. In relation to social
movements, different studies analyzed the influence of these movements on the media. Terkildsen and colleagues’ (1998) study showed that during the abortion debate in the US in the 1960s and 1970s, the media picked up several frames from social movements. Terkildsen and colleagues argue that even though the media sometimes created their own frames, most of the times strategically produced issue frames of the movements determined the news discourse. Rucht and Verhulst's (2010) study showed, in the context of the Iraq war, that social movements against the war were able to incorporate slogans, claims and frames into the news coverage. Due to the increased coverage of Pegida in German news since its foundation and based on the above mentioned arguments, I suggest an influence of Pegida's claims and frames on the news media coverage on refugees and asylum seekers. Since Pegida is an anti-immigration and anti-Islam movement with a ‘negative’ attitude towards
immigration and slogans against multiculturalism, emphasizing the abuse of the asylum system, I expect a rise of negative frames:
Hypothesis 2: I expect a rise of the negative frames about refugees in German regional newspapers after the foundation of the Pegida movement.
The bright West, the dark East?
The Pegida movement was founded in the East of Germany and even though it has small offshoots in cities in the west, it centers and attracts the highest number of participants in Dresden and Leipzig (Pfahl-Traughber, 2015), the two biggest cities in the east. Eastern Germany shows a higher level of support for right-wing extremist parties as well as problems related to extreme-right violence against minorities (Quent, 2012). A study by Decker and Brähler (2008) investigated that 32.6% of the East German population show xenophobic attitudes, more than twice as high as in Western Germany. Further, since the reunification of Germany, right-wing political parties, e.g. NPD (Nationaldemokratisch Partei Deutschlands), achieved only in Eastern Germany significant success. The NPD was elected in two state
parliaments, in one of which they are still represented today. The reasons for this difference are still a point of debate. One of the most common explanations is the aftereffect of the dictatorial past of the former GDR. A majority of the population in Eastern Germany grew up and lived under the socialist regime and thus absorbed certain political mentalities, which make them more prone to extremist positions (Quent, 2012). Furthermore, there is still a gap in the socio-economic conditions between the West and the East. The unemployment rate is higher, and the net wages lower than in Western Germany. In addition, the level of trust in democracy is lower in the East than in the West (Quent, 2012). Therefore, it is important to investigate possible differences in the framing between Eastern and Western newspapers.
RQ2: Is there a difference in framing refugees between the Eastern and Western German regional newspapers?
Methodology
To investigate how regional newspapers in Germany frame asylum seekers and refugees, I conducted a research study with a longitudinal design. Including a quantitative content analysis of articles from four different regional newspapers, this study sets out to analyze news frames. In order to deeply analyze and understand common news frames and identify relevant issue-specific frames in the German regional press, a quantitative content analysis is the best choice of method. Content analysis is said to be a more powerful method than surveys and interviews, because it is less dependent on subjective perceptions and because of its unobtrusive approach (Krippendorff, 2004). The study is further solidified by the usage of more than one coder and the verification of intercoder reliability (Boettger & Palmer, 2010). Thereby, content analyses are extremely useful, especially in framing research (Entman, 1993).
Sample
weeks. The analysis investigates the 20 weeks leading up to the first march of Pegida on the 20th October, 2014, and the same amount of time after the foundation of the protest
movement. The period of the study begins with Monday 2nd June 2014, and ends with Sunday the 8th March 2015. The content analysis was carried out on a sample of news articles from four different German regional newspapers, the Nürnberger Nachrichten (NN), Rheinische
Post (RP), Sächsische Zeitung (SZ) and Mitteldeutsche Zeitung (MZ). The German regional
press was chosen because this study set out to study differences between East and West Germany. An East-West comparison is not possible with a sample consisting of national newspapers. Furthermore, in Germany regional newspapers still have a large circulation and play an important role in citizens' lives. In 2013, regional daily newspapers reached a
readership of approximately 31 million people per day. This is 51.6% of the
German-speaking people in Germany aged 14 or above. In comparison to that, only 3.7 million people (5.3%) read a daily national newspaper every day (Pasquay, 2014). Furthermore, comparative research with content analyses has a tendency to focus on national news coverage, which is useful in areas with national importance like defense or international trade. However, topics like immigration “have a distinctly local component” (Lawlor, 2015, p. 919), since refugees and asylum seekers are also distributed to rural areas with low population. Lawlor (2015) argues that immigration constitutes an area where the regional perspective is important and relevant.
The four regional newspapers were included in the analysis based on circulation numbers as well as geographical location. In order to compare the coverage between East and West Germany, two eastern and two western newspapers were chosen. The Rheinische Post (RP) (circulation: ca. 315,000) is the second biggest regional newspaper in the West German state of North Rhine-Westphalia, which is the most populous state in Germany. Nürnberger
Germany, with a circulation number of approximately 260,000. The two eastern German newspapers are Sächsische Zeitung (SZ) and Mitteldeutsche Zeitung (MZ). SZ is from the most populous state of Eastern Germany, Saxony, and has a circulation of approximately 235,000 papers. MZ is read in the Eastern state Saxony-Anhalt and has a circulation of approximately 187,000 papers.
From these four newspapers, articles were derived through LexisNexis by applying a specific search string (see Appendix B), which contained the most relevant terms and phrases in connection to refugees and asylum seekers. The articles were further selected by the creation of an 'artificial week', which means that in every week another weekday was selected, e.g. Monday in week 1, Tuesday in week 2 etc. This method ensured that the distortion in regards of the content of the newspaper as a whole was kept to a minimum (d'Haenens & de Lange, 2001). Through pre-sampling, reader contributions and news roundups were removed from the sample since they do not reflect how professional news journalists frame news. Finally, from this pre-selection 505 articles were selected (RP n=140, NN n=115, SZ n=123, MZ n=127) and coded. The unit of analysis and coding unit was the news article. Even though the selection of the articles was derived from regional newspapers, it also included articles with national and international focus. However, the articles with local and regional focus were clearly predominant (n=401, 78,9%).
Coding Procedure
The coding of the articles was conducted by one trained coder according to a specific codebook (see Appendix C for a list of relevant variables). Firstly, the article was read completely to catch the main topic and actors of the news story. Afterwards, the article was read again paragraph by paragraph and each variable was coded in the order specified in the codebook. Thereby, general information is dealt with first, and while coding the frame items the coder possess an overall picture of the story. In order to study the complex topic of
refugees and asylum seekers in the news, the codebook consisted of a variety of different variables.
Variables. After basic introduction variables like date, medium, journalistic form, source, length and regional focus of the article, the codebook also asked for topic and actors. The topics and actors were coded using hierarchical codes with seven (topic) and nine (actor) main categories and numerous subcategories. Topic main category 300 for example dealt with the life of refugees in Germany. Within this main category, there were several
subcategories, for example accommodation (301), caring for refugees (302), integration of the refugees (303) etc., which the coder needs to choose from in order to specify the topic of the article. The same structure of coding applied for the actor of the article.
The codebook further tapped the country of origin of the refugees and the usage of emotional language or metaphors in the article, and asked for the tone of the article towards the influence of the refugees on the region. In the country of origin variable, there are 19 different countries and regions to choose from. The emotional language variables
investigated the presence of words related to emotions, which might be affected through the establishment of the anti-immigration movement. Furthermore, emotional language might also serve as predictors for episodic or thematic framing and were therefore used as control variables in a linear regression. Therefore, a total of 15 emotions were included in the
codebook (see Appendix C). Variable 16 (despair) for instance, had to be coded with Yes (1), if words like desperate, despaired etc. were explicitly mentioned in the article. The same system applied for the other emotional language variables. Even though not all variables of the codebook were necessary to answer the research questions and to verify the hypotheses, a codebook with a wide range of variables enables the study to fully investigate the coverage on refugees and asylum seekers.
measures developed by the respective authors. Every frame consisted of a different number of frame items. For example, the intruder frame (Van Gorp, 2005) consisted of four items: a) “Are refugees or asylum seekers referred to as a source of trouble or conflict?” b) “Are refugees or asylum seekers in the article connected to illegality or illegal methods?” c) “Are refugees referred to as dangerous?” and d) “Does the article state that refugees abuse the asylum system?” The codebook (see Appendix C) offers a full list and explanation of the applied frame items.
Regarding the humanitarian and bogus-refugee frame, I developed a list of several operationalization items specifically for this study. As mentioned above, literature points out to the differentiation between bogus refugees, who come because of economic benefits and genuine refugees, who flee because of war or humanitarian catastrophes. The items applied in this study build on the work from Ettinger & Udris (2009) and Nickels (2007), in which a qualitative approach was chosen, to derive the frames from news articles. Build on their descriptions, a number of operationalization items were developed to apply in this quantitative study. The bogus-refugee frame consists of five items and the humanitarian frame consists of four items (see Appendix C).
Regarding thematic and episodic framing, the operationalization items were derived from Iyengar's (1991) analysis of television news, and consisted of one item each: “Does the article tell the story with the help of one specific individual or does it focus on one specific event?” (episodic) and “Does the article give background information and put the story into wider (historical, analytical) context?”.
All the frame questions applied in this study had to be either answered with Yes (1) or No (0) and were applied to every individual newspaper article.
Intercoder Reliability
assessed on a randomly selected 50-article sample (approximately 10% of the main sample). The measure provides reliability, even if certain data is missing and regardless the number of coders or the sample size (Hayes & Krippendorff, 2007). The alpha scores were in an
acceptable range (from 0.60 to 1.00), except for a few variables, with scores reaching from 0.48 as low as 0.04 (for a full list of the scores see Appendix D). Three of the emotional language variables for example (variable 19, 20 and 21) showed low alpha scores (0.48, 0.04 and 0.45). These scores resulted from an inaccurate description of the coding rules in the codebook. The coding rules descriptions were improved, by mentioning that only if words of emotional language related to poverty, drama or mistrust are explicitly mentioned in the text, these variables can be coded with Yes (1). After the test, refinements of the initial codebook were made in order to improve the accuracy and reliability of each variable. Variable 31, regarding the tone of the article towards the influence of the refugees on the region, also showed a low alpha score. To improve the score, refinements in the codebook were made, to show that coverage about positive events, like a welcome party in an asylum center, does not mean that there is a positive influence of the refugees. Only in the case of an explicit
mentioning of positive or negative influence of the refugees, this variable is to be coded with the equivalent code. Variable 34, about the portraying of refugees as threats, also showed a low Krippendorff alpha score. In this case, the codebook was improved by defining that also in the case that refugees are being referred to as threats indirectly, for example by one of the local citizens, this variable needs to be coded with Yes (1). Variable 45 on the portrayal of refugees as economic threats, also was improved by defining more precisely which words need to be present. Words like 'financial pressure', 'burden' or 'threat' need to be explicitly mentioned in the text in order to code the variable with Yes (1).
Data Analysis
Semetko, 2010; Valkenburg, Semetko, & de Vreese, 1999), multi-item scales were formed, for each of the frames, by adding the scores of each item and dividing the total by the number of relevant items. Thus, the values of each scale range from .00 to 1.00. A high score on the 'intruder' scale for example means that the article was presenting refugees related to illegality, danger and abuse of the asylum system.
In regards to the two research questions, I conducted ANOVAs to enable mean
comparisons. Furthermore, to answer the second research question and to test hypothesis two, t-tests reveal possible mean differences between the Eastern and Western newspapers and before and after the foundation of the Pegida movement.
In order to test the first hypothesis, this study further applies hierarchical linear regression models to show how thematic and episodic is affected by positive and negative issue-specific frames as well as other control variables. I employ a series of models
successively adding new explanatory variables, thereby assessing the explanatory value of each of the models and the improvement of the model when introducing new variables. The first model (Model 1) accounts for emotional language. Emotions or the presence and absence of emotional language in an article might play a role in predicting episodic and thematic framing, since both frames come with different emotional details as well as generate different emotional reactions (Aarøe, 2011). Specifically episodic frames carry different “human interest details” (Aarøe, 2011, p. 210), which are more emotionalizing and
personalizing than pale statistics (Gross, 2008). Therefore, I control in the first model of this linear regression for the impact of a selection of emotional language on episodic framing, namely ‘anger’, 'despair', ‘gratitude’ and 'joy'. Previous studies (Aarøe, 2011; Gross, 2008; Nabi, 2003) have connected different emotions like anger, fear, pity, or disgust with episodic framing. I have chosen two positive and two negative emotions, since this study also
dummy variables related to the topic and actors in the news story. These variables include the actor 'politician', the topic 'finance', 'accommodation' and 'escape', which is focused on the personal history of the refugees and their escape from their home country to Germany. Finally, the third model (Model 3) then incorporates the six positive and negative issue-specific frames to test their influence on thematic and episodic framing.
Results
To give an overview of the coverage on refugees and asylum seekers in all
newspapers, several general observations will be presented first. The most dominant topic within all the articles from all four newspapers was by far 'accommodation'. 54% (n=273) of all the articles dealt with issues related to the accommodation of the refugees in their new place of residence. Questions often asked within these stories were, for example, whether new arrivals should be accommodated in big residential homes or in small, private apartments. This topic was followed by the topic 'statistics and numbers about refugees' (33.63%, n=170) and by 'asylum policy' (24.75%, n=125). The most represented groups of actors were civil servants, working in city halls or public offices, who were present in 40.79 % (n=206) of the articles. The second biggest group of actors was local politicians, who were present in 33.46% (n=169) of the articles, followed by individuals who support refugees through voluntary service in nonprofit organizations (26.33%, n=133). In comparison to that, refugees whose names were mentioned in the articles appeared in only 11% of the stories (n=60).
Framing Coverage
The first research question of this study concerned the proportion of positive and negative issue-specific frames used in the coverage of refugees in the four selected newspapers. Regarding the overall presence of issue-specific frames, 39% (n=198) of all articles contained no frame, meaning that all operationalization items of all six issue-specific
frames were coded 0. Table 2 shows the mean visibility and the standard deviation of all eight applied frames in this study. As can be seen, the values ranged from M = .002 (Islam-as-a-threat frame) to M = .16 (Humanitarian frame). As table 2 shows, all three positive frames had higher mean visibility than the negative frames. Thus, in my sample, refugees and asylum seekers were more frequently framed with positive than with negative frames.
< TABLE 2 ABOUT HERE >
Figure 1shows the mean development over the individual months, to see whether the positive frames exceed the negative frames during the whole period. Figure 1 shows that the positive frames were almost always more prominent than the negative frames, with a short exception in October when the intruder frame exceeds the level of the victim and
multicultural frame.
< FIGURE 1 ABOUT HERE > Predicting Episodic and Thematic Framing
< TABLE 3 ABOUT HERE >
In H1, I hypothesized that episodic framing is positively associated with positive frames, and thematic framing is positively associated with negative frames. In regards to this hypothesis, I conducted a linear regression to investigate which positive and negative frames predict episodic and thematic framing respectively. Table 3 shows the prediction of different factors on episodic and thematic framing in three different models per frame. Firstly, model 1.1 starts with incorporating several variables related to emotion, to test the impact of
emotional language on episodic framing. As can be seen, all emotion variables, except anger, have a strongly significant positive influence on episodic framing, meaning that if a news stories contains emotional language in relation to despair, gratitude or joy, this story is also more likely framed in an episodic way. However, the explanatory power of the model is low, explaining only 7.8% of the variation in the dependent variable.
In the second model (Table 3, Model 1.2), I add several dummy variables concerning topic and actors of the article, which as well might influence the degree of episodic framing. As table 3 illustrates, these variables all show a significant negative influence on episodic framing, with the exception of the topic 'escape'. This means that articles about finance or accommodation or stories including politicians as actors are significantly less often framed in an episodic way. The explanatory power of model 2 rises significantly by 8.2% to 16.0%.
In the third model (Table 3, Model 1.3), and to provide a final test for the hypothesis, I incorporate the six positive and negative issue-specific frames. As expected, Table 3 shows that all positive issue-specific frames show highly significant positive influence on episodic framing, with the positive multicultural frame being the strongest predictor. This means that in cases in which refugees are framed in a positive way, the news stories also tend to be framed more often in an episodic way. However, contrary to my expectations, the negative intruder frame also shows a significant positive correlation to episodic framing, meaning that news stories which frame refugees as intruders or potential threats, are more often framed in an episodic way. Table 3 shows that the issue-specific frames, especially the positive
multicultural and victim frame, are considerably the strongest indicators of thematic framing, since the explanatory power of the model significantly rises from 16.0% to 39.2%.
In a next step, the same models were further applied to thematic framing as the dependent variable. As can be seen in table 3, the emotional language variables 'joy' and ‘gratitude’ have a significant negative influence on predicting thematic framing, meaning that articles, which contain words related to joy or gratitude, are less likely thematically framed. The other emotional language variables show no substantial impact on thematic framing. Regarding the explanatory power, model 2.1 only explains 2.9% of the variation in the dependent variable.
In the second model (Table 3, Model 2.2), the dummy variables are being added again. As can be seen, all of the newly added variables have a significant positive impact on the dependent variable, while slightly reducing the influence of the significant variable from model 2.1. The strongest predictor of thematic framing appear to be 'finance', meaning that news stories about the costs of refugees are more often framed in a thematic way. The explanatory power of the model rises significantly to 11.3%.
In the last step, the positive and negative issue-specific frames are being added to the third model (Table 3, Model 2.3). As can be seen, the negative bogus-refugee frame shows significant positive influence on predicting thematic framing, meaning that the higher the level of the bogus-refugee frame, the higher the level of thematic framing. However, not in line with the expectations, the other two negative frames, namely the intruder and the Islam-as-a-threat-frame, are not substantially related to thematic framing. Regarding the positive frames, only the multicultural frame has a significant negative influence on thematic framing, meaning that stories that emphasize respect and exchange between cultures are less likely to be framed in a thematic way. Table 3 shows that the issue-specific frames, foremost the negative bogus-refugee frame and the positive multicultural frame, are the most influential predictors for thematic framing, by raising the explanatory power of model 2.3 significantly to 20.6%. In sum, the first hypothesis is confirmed.
The influence of Pegida and the difference East and West
I now turn to the second hypothesis in which I expected a higher level of negative frames after the foundation of Pegida. An independent-samples T-Test compared the means of the frames before and after the first demonstration of the anti-immigration movement. Regarding figure 2, it becomes evident, in line with my expectations, that all negative frames experience a rise after the foundation of Pegida. The intruder frame raises from M=.056 (SD=.172) to M=.065 (SD=.189); t(503)=-.546. The bogus-refugee frame increases from
M=.025 (SD=.090) to M=.041 (SD=.136); t(503)=-1.54. However, both frames do not show
any significant difference (p > .05). The increase of the Islam-as-a-threat-frame is not significant either, from not existent at all before Pegida to be present in exactly one article after Pegida (M=.0038, SD=.061; t(503)=-.967, p=.318).
< FIGURE 2 ABOUT HERE >
Regarding the positive frames, the results find significant and meaningful change before and after Pegida. As figure 2 shows, and in line with my expectations, the
humanitarian frame decreases from M=.182 (SD=.274) to M=.141 (SD=.233), with a p-value approaching significance; t(503)=1.79, p < .10. However, contrary to our expectations, the positive multicultural frame experiences a significant increase from M=.076 (SD=.197) to
M=.164 (SD=.281); t(503)=-4.05, p < .05. The victim frame shows the same mean visibility
before and after (M=.117, SD=.252; t(503)=.000, p=1).
Regarding the mean development of the individual frames by month, there are differences recognizable, however these differences are not systematic and in most cases not statistically significant. A oneway ANOVA revealed, that the multicultural frame (F(9, 495)=3.576) for instance, reaches a peak in December (M=.213, SD=.314) which is significantly different to the values in October (p=.023) and June (p=.011). The bogus-refugee frame (F(9, 495)=2.708) also shows certain significant differences by month. It reaches the highest mean visibility in the first two months of 2015, January (M=.074,
SD=.174) and February (M=.080, SD=.199). The mean values from February are significant
different to the values of the last two months in 2014, December (p=.022) and November (p=.037). However, alternative explanations for these developments will be given in the discussion part. In sum, H2 is not supported.
The last research question concerned the differences in the usage of negative and positive frames between Eastern and Western regional newspapers. Also here, an
independent-samples t-test revealed contradictory results. Regarding episodic and thematic framing, the comparison between East and West shows that the Western newspapers frame news stories on refugees to a slightly, non-significant higher extent in an episodic way (M=.30, SD=.460) compared to the East (M=.26, SD=.437); t(503)=1.15, p=.251. In line with this, the East shows a significant higher level of thematic framing (M=.59, SD=.493)
compared to the West (M=.48, SD=.501); t(503)=-2.38, p=.017, as can be seen in figure 3. < FIGURE 3 ABOUT HERE >
The negative intruder frame is significantly higher in the East, meaning that the two selected Eastern German newspapers (M=.086, SD=.217) frame refugees more often as intruders and potential hazards than Western German newspapers (M=.036, SD=.130);
t(503)=-3.12, p=.002. However, as can be seen in figure 3, the negative bogus-refugee frame
shows significantly higher mean visibility in the two Western newspapers (M=.051,
SD=.016) than in the Eastern newspapers (M=.016, SD=.078); t(503)=3.41, p=.001. The
Islam-as-a-threat-frame does not play an significant role, since it only occurs in one article;
t(503)=-1.01, p=.313.
Regarding the positive frames, figure 3 shows a significant higher level of the humanitarian frame in Western Newspapers (M=.202, SD=273) than in Eastern Newspapers (M=.119, SD=.225); t(503)=3.75, p=.000. The victim frame as well shows higher mean visibility in the West (M=.132, SD=.257) than in the East (M=.102, SD=.247); t(503)=1.30, however with an insignificant p-value score (p=.192). The mean visibility of the multicultural frame between West (M=.121, SD=.250) and East (M=.122, SD=.246) is approximately on the same level; t(503)=-.050, p=960.
< FIGURE 4 ABOUT HERE >
To further research the differences between Eastern and Western regional
the frames between the four individual newspapers. The test revealed significant differences in terms of the intruder frame (F(3, 501)= 18.151). Figure 4 illustrates the great distance in mean visibility between the Sächsische Zeitung (East) (M=.159, SD=.280) and the other three newspapers (NN: M=.034, SD=.131; RP: M=.037, SD=.130; MZ: M=.015, SD=.087) with a significance at the p < .001 level. In terms of the positive humanitarian frame (F(3,501)= 6.990), the Western newspaper Nürnberger Nachrichten shows the highest mean visibility (M=.239, SD=.291) and significance to the two Eastern newspapers, Sächsische Zeitung (M=.094, SD=.211, p=.000) and Mitteldeutsche Zeitung (M=.142, SD=.236, p=.015), which is illustrated in figure 4. Interestingly, the same applies for the negative bogus-refugee frame (F(3,501)=4.305). NN shows the highest mean visibility (M=.051, SD=.158) with a strong significant difference to SZ (M=.019, SD=.097) and MZ (M=.012, SD=.054), presented in figure 4. In sum, the t-test and the oneway ANOVA did not reveal any clear distinction in the usage of negative and positive frames between the East and the West. So, my answer to the second research question is that there is no systematic difference between Eastern and Western newspapers in framing refugees.
Discussion
This study presented the results of a content analysis of the news framing of refugees and asylum seekers in German regional newspapers. This is relevant, since the way the media frame refugee has the power to influence anti-immigration attitudes (Freeman et al., 2013).
First, I found that regional newspapers in Germany overall tend to apply positive frames to a higher extent in the coverage on refugees and asylum seekers than negative frames. This is a strong argument against a number of recent studies which argued that immigration has increasingly been covered through a negative narrative of ‘illegality’, reflecting mostly opinions of leading politicians and governments (e.g. Figenschou & Thorbjørnsrud, 2015; Benson, 2013; Gabrielatos & Baker, 2008; Horsti, 2007; Kim,
Carvalho, Davis, & Mullins, 2011; Suro, 2011). The results of this study hint in the opposite direction and are more in line with previous research on framing of refugees, which argues in favor of a predominance of positive and more liberal frames (e.g. Steimel 2010, van Gorp 2005). Immigration is a constant topic in the public discourse in Germany. Traditionally German identity has been defined by ethnicity and blood-lineage (Brubaker, 1992). However, Germany understands itself slowly but steadily as an immigration country, following the model of countries like Canada or Australia (Bauder, 2007). In this development, economic factors might play a role. The German population is aging, with one of the lowest birth rates worldwide (Meier-Braun, 2013). Therefore, Germany is dependent on refugees for future economic growth. Bauder (2007) suggests that there is a paradigm shift at the discursive level, which shifts from the closed, ethnic conception of German nationality towards an identity that emphasizes economic achievement and welfare to which immigration can contribute. The beneficial side of immigration is more and more realized in the field of politics. Meier-Braun (2013) argues that there have been more actions for better integration of immigrants and refugees in the last ten years than in the last four decades. Immigration is considered more and more as beneficial for Germany and I have reason to believe, that this shift might be partially responsible for the positive framing.
What is striking is that the Islam-as-a-threat frame only appeared in one news article, and is therefore basically non-existent in the coverage of refugees in German regional newspapers in the selected period of time. Its counter-frame, however, the multicultural frame, was the second most dominant frame in the analyzed coverage. This finding is in sharp contrast to studies from the Netherlands (Roggeband & Vliegenthart, 2007) or Australia (Sulaiman-Hill et al., 2011), which both find an increase of the Islam-as-a-threat frame or negative framing of Muslims after 2001. Often media discourse is influenced by the political discourse (Vliegenthart & Roggeband, 2007). Therefore, the political discourse in
Germany needs to be considered as well in regards to this result. I have reason to believe, that the political discourse in Germany is different than in other European countries. As
mentioned above, the German political scene does not possess a prominent right wing populist like Geert Wilders in the Netherlands, Jörg Haider in Austria or Jean-Marie Le Pen in France. Also, Germany did not experience a major Islamist terroristic attack. Therefore, the portrayal of Islam as a threat might be not a constant part in the public discourse as in other countries. I suggest, that the public discourse in Germany about Islam focuses on other issues, which might have higher importance for the German public. One of the main points in Germany for example is the headscarf (Bielefeldt, 2008; Schiffer, 2005). Germany’s self-perception as a Christian country embedded in Christian and Jewish culture is still prevalent (Schiffer, 2005). The ‘headscarf-debate’ began when a Muslim teacher wanted to wear a headscarf in class, in 1998. A public discourse started, which comes up regularly ever since. Critics of the headscarf did not just see Christian values under threat, but also the hard-fought emancipatory achievements and gender equality (Bielefeldt, 2008). A similar public
discourse rose after the ‘honor killing’ of a young Turkish girl in Berlin, who rejected her forced marriage. Therefore, I have reason to believe, that the absence of the connection between Islam and terrorism in my sample is due to the fact, that public discourse is focused on different issues in relation to Islam.
However, a study by Bauder (2007) showed that, Islamic terrorism and immigration is connected in German media as well, however, mostly in relation to external events, for instance, terroristic attacks like the bombing in Madrid. However, these are only short phases where media connects immigration with terrorism and which are focused on national media (Bauder, 2007). National media discusses topics in a broader and more international context (Lawlor, 2015). Therefore, I suggest that the results might be different if a sample, including national newspapers, would have been chosen. Future studies should take this finding as a
