Turkish News Coverage of Kurds: The Blurring Line between the Left and Right-wing Newspapers
Merve Erdogan Student ID: 10602577
‘Master’s Thesis’
Submitted in Partial Fulfilment of the Requirements for the degree of (Research) Master of Science in Communication Science
Supervisor’s name: Penny Sheets Thibaut Date of completion: January 29, 2016
Abstract
Media, as an ideological state apparatus, plays a critical role in depicting various groups in societies. The renderings of ethnic groups, races, and religions are often considered contentious disputes. For this reason, media portrayals of - commonly - subordinate collectives receive immense attention from Communication Science along with from
Political and Social Sciences academics. As one sui-generis example, this study scrutinizes the media representation of Kurds, the second largest ethnic group living in Turkey.
Although this subject has been well studied through a number of qualitative research methodologies, the literature scarcely covers an examination with quantitative research methodology. This subsequently undermines our ability to reach a systematic and replicable outcome. Therefore, for this analysis, a quantitative content analysis was conducted. 704 news articles from two opinion-leading Turkish print news outlets, Hürriyet and Sabah, from divergent ideological pillars, were coded to compare if they portray Kurds in a positive or a negative fashion. The findings suggest that notwithstanding these
newspapers’ ideological affiliations, the news coverage of Kurds does not truly show a discrepancy from one newspaper to another. This suggests that, while it claims to be a pluralist system, the Turkish media system seems rather centralized concerning the portrayal of Kurds, regardless of the true political character of the newspapers.
Nonetheless, the overall balance of positive/negative associations made in discussing Kurds in the news articles differs through the years 2010-2014. While this fluctuation in coverage cannot be attributed to newspaper ideology nor to major socio-political events, such as general elections and social unrest, it can be an outcome of a changing media landscape in Turkey. Thus, this study suggests that this complex issue requires much more scholarly attention in the future.
Key words: Kurdish, Turkish, newspaper ideology, ethnic groups, media portrayals, quantitative content analysis
Turkish News Coverage of Kurds: The Blurring Line between the Left and Right-wing Newspapers
How certain ethnic groups in different countries are covered in different media outlets prompts scientific curiosity among numerous researchers from several scientific disciplines. Following much of this research, and grounded in Hallin and Mancini’s media systems framework (2004), this study focuses on a sui-generis case - Kurds, who constitute the second largest ethnic group living in Turkey. It is a critical opportunity for this study to understand the nature of certain news stories’ content that particularly involve a certain ethnic group, as well as how these news stories are framed. Previous studies pointed out that the Kurds were not visible in the mainstream state and media discourses (Somer, 2005b; Sezgin & Wall, 2005; Çeliker, 2009; Yeğen, 2007). Through socio-political changes over time (e.g. the military coup in 1980), the political discourse has transformed and consequently the visibility of Kurds in media texts has risen (Somer, 2005b).
However, the portrayal of Kurds in media texts remained insufficiently studied (Demir & Zeydanlıoğlu, 2010). Moreover, these studies analysing (the media portrayals of) Kurds in particular, mostly out of the communication literature, employ qualitative research
methodologies such as discourse analysis, and conduct analyses on one media outlet only (Somer, 2004, 2005b; Sezgin & Wall, 2005). Hence, they hardly offer an objective,
systematic and replicable analysis of (the media portrayal) of Kurds. Therefore, this study aims to distinguish itself from existing scholarship on this topic by investigating if the Kurds are portrayed with positive or negative vocabularies in Turkish print media. It employs a quantitative research methodology, and offers a comparative approach with the inclusion of two ideologically different newspapers in the analyses. It tries to find out whether Kurds are portrayed differently in different newspapers and if the trends in portraying Kurds change over time and around certain times, for instance before or after elections or an event/incident involving Kurds. Two opinion-leading newspapers, Hürriyet – known to be an
anti-government/ Republican/secular newspaper – and Sabah – known as a
opposing political ideologies embraced by different media outlets could be an explanatory factor for how news stories are produced and published/broadcasted.
This research has academic relevance due to its contribution to, especially, the communication literature, by providing an appropriate example for scholars interested in media portrayals of ethnic groups in other countries that have ethnic struggles or had them in the past. Moreover, the findings carry an immense socio-political importance since it has repercussion for the ongoing Kurdish-Turkish armed conflict. Understanding the nature of the Turkish media system as well as the dominant state and media discourses applied when referring to Kurds may provide valuable insights for Turkey’s future ethno-politics. It may also help Turkey to define its regional role in the Middle East, increasingly important due to the war in Syria causing many (Kurdish) refugees to cross-pass the Turkish-Syrian border.
Theoretical Background
Media, Society, and the Question of Representation
Media plays a crucial societal function in helping the public to understand and
interpret the world: how it is organized, structured and changed. To fulfil this function, media content usually carries hidden messages. McQuail (2010) argues that media content carries an ideology which was “decoded” in the messages of mass media (p.349). In other words, the message of the mass media is incorporated in written text or spoken words. Through these hidden ideological messages, “media conglomerates can help shape our ideologies” about certain topics, such as ethnic group perceptions (Berger, 2007, p.209; 2011, p.93). These messages, ideas and images can help the public make sense of much of their everyday experiences by simply applying primary frameworks (Goffman, 1974). Media may not be a mere observer, nor a mirror of situations, but mostly act as a meaning-creator; “a strong agent of ‘attitudinal’ effects dependent on cognitive schemas” (Panayırcı & İşeri, 2014, p.65; Scheufele & Tewksbury, 2007, p.12). Aligned with this argument, Shoemaker and Reese (1996) make a definition of framing - as a macro-construct – being modes of presentation that communicators use to present information in a particular manner which would resonate with their audiences’ existing schemas. To put differently, the media tend to
selectively choose familiar cultural themes that would go along with audiences’ structured mind maps composed of their preconceived ideas and thoughts.
Having this vital power of meaning creation in its hand, media tend to be selective in coverage, especially on controversial issues. As the most commonly used definition of framing by Entman features, the media tend to select some aspects of reality and render them more salient. In this way, it promotes a problem definition, as well as evaluations and interpretations of that particular problem (Entman, 2007). Entman (2007) further argues that framing can be used in the exercise of political power. Kinder and Sanders (1990) suggest that frames operate as “devices embedded in political discourse, invented and employed by political elites, often with an eye toward advancing their own interests or ideologies and intended to make favourable interpretations” (p.74). Therefore, one can claim that this media strategy is much needed by politics, especially to persuade the public to adopt a particular message (Panayırcı & İşeri, 2014).
Political elites as well as public opinion makers “tend to manipulate the media to disseminate news in a manner that fosters their legitimizing discourse” and this, in turn, facilitates to perpetuate their legitimizing discourses (Panayırcı & İşeri, 2014, p.65). In other words, the ruling elite has a controlling function over the information that the public receives, in order to legitimize and maintain the power structures in the society (McQuail, 2010). Although journalistic professionalism necessitates an editorial autonomy and numerous journalists, mostly in northern (corporatist media model) countries where political interests are not intertwined in news production, claim to rejoice full journalistic autonomy; media in other models, that is instrumentalized by politics or business, can still be observed as reproducing hierarchies of power, especially in the domains of ethnicity, class, race and gender (Hallin & Mancini, 2004). This would also imply that journalists might incorporate bias in their language referring to these subordinate groups.
In Marxist tradition, argue Shoemaker and Reese, the media is discussed as having relative autonomy (1996). The ruling class dominates a hegemony over media institutions producing a cohesive ideology. This, in the end, underlies the relationship between the
dominant and subordinate groups in the society. In other words, the power relations in the society are reflected with underrepresentation or stereotypical portrayals of subordinate groups in the society. For instance, women are usually depicted as inferior figures or as victims of violence; whereas men are depicted as equated with power or as victimizers in television programmes (Shoemaker & Reese, 1996). Another exemplification is that blacks and Latinos are portrayed as law breakers or perpetrators of violent crime in the news whereas whites are represented as victims of crime or as law-enforcers. These examples can explain how media can cultivate certain stereotypes. Therefore, it is crucial to
comprehend how racial or ethnic portrayals are reproduced, and how they are reinforcing biases against these particular groups (Entman & Gross, 2008).
Kurdish Issue in Turkish Context
The Kurds, as one example of a subordinate group, are one large, stateless and territorially divided national group. They are dispersed around a large geographical area around Turkey as well as its neighbouring states Iraq, Syria and Iran (Romano, 2002; Çeliker, 2009). Kurds constitute, after the Turks, the second largest ethnic group in Turkey (Sezgin & Wall, 2005). Under the rule of the Ottoman Empire, Kurds together with other groups (Arabs, Greeks, Circassians and many more) were encouraged to live peacefully under the umbrella term of “ümmet” – “an institutionalized pluralism through the millet
(nation) system” (Van Bruinessen, 1998, p.50).This “ideal of cultural mosaic” ceased to exist with the fall of the empire. Having established the Turkish nation-state around the ideals of one nation and one language, “the military-bureaucratic intelligentsia of modern Turkey” enforced ethnic group suppression (Al, 2015, p.95). Kurds were prohibited to publish or broadcast in the Kurdish language due to Turkey’s ban on broadcasting and publishing in any other language different than Turkish1 (Romano, 2002; Somer, 2002; Çeliker, 2009). This ban aimed to prevent ethnic separatism (Somer, 2004).
1
The prohibition on publishing and brodcasting in Kurdish language was abolished in 2004.
Despite of the emergence of a strong Kurdish ethnic consciousness through the years, the Kurdish political movement was largely impacted with the revival of Islamic parties following the years after the fourth military coup of the Republic of Turkey in 1980. These parties aimed to blend Kurds into the “ümmet” system, in other words, into a mono-religious Turkish identity, as it was done in the Ottoman Empire rule, whereas the military tried to solve the Kurdish “problem” by applying force (Al, 2015, p.95). These assimilation
techniques reflected themselves both in state and media discourses. Until the 1990s, using the word “Kurd” to denote an ethnic group living in Turkey was considered a taboo, a sign of breach to the country (Romano, 2002; Somer, 2002). People who use the word “Kurd” were mostly labelled as being separatists (Somer, 2002, p.86).
Due to the state oppression of the Kurdish identity, the Kurds were invisible in both mainstream media and state discourse (Sezgin & Wall, 2005; Çeliker, 2009; Yeğen, 2007). It is only possible to gather information from Turkish media which, Çeliker (2009) argues, applies a rather denigratory and exclusionary discourse by (re)producing negative stereotypes of bandits, terrorists or smugglers, or victims belonging to an inferior culture (p.92). This consequently reinforced the identification of Kurds as the “other” (Çeliker, 2009, p.92). Moreover, this negative representation of the “other” prolonged the ongoing conflict (Iskander, 2012). After 1991, the discourse was rather liberalized and the word “Kurd” became more visible in the Turkish media due to rising consciousness and the politicization of the Kurdish identity (Somer, 2002, 2004; Romano, 2002).
The 2000s witnessed major improvements in Turkey’s domestic and foreign politics due to the European Union’s pressure on democratization as a membership conditionality (Ergin, 2014; Al, 2015). That is, if Turkey wishes to obtain membership status within the EU, the EU has instituted regulations aimed to achieve certain standards with regard to
to publish and broadcast in Kurdish. There are television channels – allegedly linked to Kurdistan Worker’s Party PKK - such as MEDTV (UK and Belgium) and ROJTV (Denmark) available to everyone via satellite television channels and now via Internet, too (Çeliker, 2009).
fundamental rights of cultural minorities. This includes recognition of the Roman people, more religious freedom for Alevi people – including obtaining a right to fulfil religious practices in Alevi cemevis other than mosques – as well as to allow Kurds to speak, write, publish/broadcast and educate in the Kurdish language. These EU requirements
necessitated the withdrawal of the military from intervening in politics, which put an end to further coup d’états while opening up political and societal improvements. More specifically, the military was required to stop “handling” the Kurdish “problem”. This “problem” needed to be solved with diplomatic techniques, which resulted in re-framing the so-called “Kurdish Problem” by the Turkish political and media elites in both state and media discourse, mostly concerning cultural rights such as education in the mother tongue (Al, 2015).
The re-framing started with the removal of the restriction on Kurdish language by 2004. Further, a television channel broadcasting in the Kurdish language was established by the Justice and Development Party (JDP) government. This facilitated a rather liberalized discourse in which the word “Kurd” could be used without fear of being labelled as a “separatist”. In 2009, JDP initiated a comprehensive democracy package called the Democratic Initiative. One major chapter of this initiative is dedicated to solve the “Kurdish Issue.” This chapter aimed to end the Turkish-Kurdish conflict by improving the societal conditions of this cultural & linguistic minority living in Turkey. Somer (2005a) argues that all these historical changes in the Turkish political landscape compelled the Turkish state and media actors to develop new discursive strategies. Therefore, it is necessary to provide a more recent and systematic overview of how the Kurds are represented in Turkish media. This is an area insufficiently studied, particularly with quantitative techniques (the ones existing either offer only limited aspects of the issue by studying only one newspaper, or do not include recent analyses).
Turkish Media System
Before delving into the Turkish media framing of Kurds, it is crucial to comprehend the nature of the Turkish media. Turkish media can be described, on the surface, as a pluralist system according to Hallin and Mancini’s media systems models (Hallin & Mancini,
2004). This means that media outlets are mostly privately owned. Each of these privately owned media outlets “reflect distinct political orientations in their news and current affairs reporting” (Çakmur & Kaya, 2010, p.522). The one-to-one connection between media and politics is particularly strong in Turkey (Çakmur & Kaya, 2010). This closeness is known to be political parallelism, as defined by Hallin and Mancini, and it has not decreased despite the rise of commercialization (Çakmur & Kaya, 2010). By the same token, the increase in the number of private television channels by 1990s and onwards did not affect the increase in the newspaper readership, of national newspapers, in Turkey (Çakmur & Kaya, 2010). Nonetheless, media ownership became rather centralized by time, which implies that a few media corporations have dominance over others.2
Van Atteveldt, Kleinnijenhuis and Ruigrok (2008) hypothesize that newspapers tend to cite and portray politicians belonging to their traditional pillar more often. Unlike the Dutch scenario they analysed, Turkish newspapers tend to proclaim an open allegiance to a certain party (Çakmur & Kaya, 2010). Therefore, it is possible to expect that the political character of the media organization – apart from the medium type, content nature, the formation, and presentation of the message – has a major impact on the content of the message conveyed to the public.
Çakmur and Kaya (2010) argue that Turkish journalists claim that they adopt a neutral model of journalism, but many studies disclose that these journalists usually tend to align with the interests of the newspaper owners. They explain the sharp media division through the Sabah newspaper example: Sabah newspaper, one of the opinion leading newspapers published in Turkey, belongs to a conservative/right media pillar which is chiefly controlled by the governing Justice and Development Party, and thus disseminates their viewpoints (Çakmur & Kaya, 2010). In other words, it tends to reflect the current ruling elites’
2 For instance, Doğan Group is the largest media group in Turkey today and the Hürriyet newspaper,
one of the newspapers that was analysed in the study, belongs to this media group (Çakmur & Kaya, 2010, p.531).
conservative-religious discourses, especially with regards to Kurdish issues (Al, 2015). On the other hand, the biggest daily of Turkey, Hürriyet, another opinion leading newspaper, belongs to Doğan Media Group that “voices criticisms against the government” (Çakmur & Kaya, 2010, p.532). For that regard, it is important to investigate if Turkish (print) media that belong to different political pillars apply different discourses in portraying Kurds.
Based on their findings of a comprehensive content analysis, Somer and Liaras (2010) claim that religious-conservative and secular newspapers differ from each other with regards to national identity and ethnic diversity. They concluded that the news coverage of religious-conservative newspapers, with close-ties to the ruling JDP, tends to adopt a rather open/positive outlook whereas the secular newspapers, that are known to support
Republican positions, tend to be rather negative. This could again be explained by the high level of political parallelism. Based on their conclusion, it is hypothesized that the Sabah newspaper will frame news more positively by portraying Kurds with more positive word use than Hürriyet. For instance Hürriyet, known to be a centre-left anti-government newspaper with its defensive-nationalist perspective, would frame news about Kurds rather negatively by portraying Kurds with more negative associations. It takes an opposition stand and so tends to “be less likely to support Kurdish ethno-nationalist movements and groups” (Sarıgil, 2009, p.533). Hürriyet might portray Kurds as a threat to the nation’s unity; whereas Sabah, known to be a centre-right newspaper having strong ties to the ruling JDP and with its rather liberal perspective, would tend to reflect the JDP’s view of Kurds and so be more open about the ethnic diversity and the inclusion of Kurds by employing a more positive word choice. For this reason, the first hypothesis would be proposing that newspapers’ political character – left or right wing ideology – could have an impact on how news stories about Kurds are framed – or if Kurds were portrayed positively or negatively.
Hypothesis I: News coverage of Kurds will tend to be more negative in the newspaper Hürriyet than in the newspaper Sabah.
Jonkman & Verhoeven (2013) discuss that frames are constructed through words. Political elites tend to “display the ability to select a certain vocabulary over another” in accordance with their political predispositions (Simon & Jerit, 2007, p.255). Moreover, framing literature seems to undermine the importance of understanding the nature of frames (Van Gorp, 2005, p. 486). For instance, valence frames, as defined by De Vreese and Schuck (2006), are strong indicators that possibly carry good or bad/positive or negative elements (p.6). These frames are very beneficial to “evaluate political issues or situations in either positive or negative terms” (De Vreese & Schuck, 2006, p.6). For that regard, the sub-hypothesis asks if news stories about Kurds employed these valence frames carrying either positive or negative elements in portraying Kurds in Turkish print media:
Hypothesis Ia: The overall balance of positive/negative adjectives will be more positive in Sabah than in Hürriyet.
Somer (2005b) claims certain association of Kurds can be made either from a security-oriented (i.e. terrorism) approach or a non-security oriented approach. The security-perspective discusses Kurds as a threat to the unity of the nation-state, mostly in relation to the terrorist organization Kurdish Workers’ Party (PKK) (Al, 2015).Somer
(2005a) argues that requesting Kurdish cultural rights would affect national unity. According to Van Bruinessen (1998), this security association could explain the causal relationship between armed conflict and ethnic identity realization. To put it differently, “the armed conflict has increased the Kurds' awareness of their distinct ethnic identity and caused an acute sense of being regarded as second-class citizens at best” (Bruinessen, 1998, p.50). This would, in return, have an impact on how Kurds are portrayed.
Secondly, Kurds might be discussed from a non-security perspective. The possible associations could be with cultural minority rights. This issue can be analysed on two axes. The first is the European Union’s membership conditionality. Secondly, the Kurdish issue is particularly represented in many cases as an issue of language – broadcasting, publishing and teaching in Kurdish language. Servet Mutlu (1996) claims that “the language is the
principal marker of Kurdishness” (p.519). For that regard, it will be reasonable to expect that newspaper texts may, at a large extent, cover the Kurdish issue in relation to language rights issues such as education in the mother tongue. (Having this information in mind, the
“democratic/cultural rights” item is created in the codebook). Similarly, if any text refers to a person (e.g. “an ethnic Kurd” or “ethnically Kurdish”), Mutlu (1996) suggests it is possible to expect that there is a reference to the language of Zaza or Kırmançi that is spoken.
Another crucial non-security association is relating Kurds to Islam. The pro-Islamic discourse, for instance, would emphasize the connection between Turks and Kurds with a “Muslim Brother” association by implying an end to separatist claims, whereas the secular-republican discourse would undermine this association (Sarıgil, 2010, p.543). All these discussions lead to the second sub-hypothesis of the study:
Hypothesis Ib: The overall balance of positive/negative associations in news articles will be more positive in Sabah than in Hürriyet.
Change in Coverage over Time
The second part of this study is rather exploratory. It aims to find out if the coverage of Kurds changes over time. Somer (2004) discusses that under periods of uncertain changes, for instance the democratization steps regarding Kurdish language liberalization, “people’s self-identities become prone to shift” (p.237). It is possible to observe that the identities might become more responsive to changing state policies and discourses. Moreover, the influences of these changes can be captured in media discourse as well. Thus, it is important to find out if linguistic categories or associations describing the Kurds change over time. Despite the fact that it could be difficult to examine potential shifts around particular incidents, Romano (2002) claims that some particular incidents could create a greater sense of ethnic consciousness and therefore it is possible to expect differences in coverage over time. As discussed before, Sabah newspaper, with its liberal-national
standing resulting in more a favourable approach toward Kurds, is expected to have a rather positive coverage of Kurds over time. Yet, if any event/incident occurred, involving Kurds as the source of conflict or initiating unrest in society, Sabah newspaper is expected to take a
rather defensive-nationalist standing similar to Hürriyet newspaper, and may tend to cease emphasizing religious unity and put more emphasis on the national unity of the country instead (shifting from positive to negative). Hürriyet, on the other hand, is not expected to change its approach towards Kurds after an event/incident due to its already held defensive-national standing. Therefore, taking differences in newspapers’ political characters into account, the second hypothesis follows as:
Hypothesis II: The overall use of positive/negative adjectives and associations of Kurds will differ before and after (non)conflictual times, through the years, in Sabah while it will stay constant in Hürriyet.
Methodology
A systematic analysis of national print news media in portraying the Kurds is crucial to gain insights about recent media discourse. For that purpose, a quantitative content analysis was carried out on a sample of articles from two national Turkish newspapers. Hürriyet and Sabah were chosen first due to their apart ideological standing with regards to Kurds and issues involving Kurds. Hürriyet is widely chosen for analyses among other scholars, due to it having the highest circulation and advertising revenues among Turkish newspapers (Sezgin & Wall, 2005). Sabah did not receive that much attention due to its frequent management changes and volatile circulation numbers. This study, therefore, could be considered as the first systematic analysis of Sabah newspaper in this regard. Secondly, these newspapers are among the most circulated in Turkey and therefore highly influential. Hürriyet is the second highest circulated newspaper (circulation by December 2015: 356,728), whereas Sabah is the fourth highest circulated newspaper (circulation by December 2015: 307,054) by the end of 2015 (medyaradar.com, 2015). Therefore the inclusion of these newspapers contributes to the representation of Turkish mainstream newspapers. Lastly, from practical reasons, these newspapers are the only newspapers with effectively working online archives to reach the sample.
The sample was collected from the online archives of these newspapers among articles published between December 29, 2010 and May 31, 2014 in order to have a rather
recent andcontinuous period of coverage. The starting and ending dates were determined randomly. Due to the idea that a high level of political parallelism goes hand in hand with low levels of journalistic professionalization, it is considered that journalists of Turkish newspapers would not present full journalistic autonomy in their articles (Çakmur & Kaya, 2010). For that reason, editorial/opinion pieces were excluded from the analysis to toss out the potentially biased journalist effect in general but focusing more on overall newspaper ideology. Moreover, there has been found no earlier studies that provide a significant correlation between media discourse and editorials (Simon & Jerit, 2007).
The sample initially includes all newspaper articles that contain the word “Kurdish” (Kürt in Turkish) in either the headline or the full text. However, to avoid contextual
differences,3 only “relevant” articles were collected —those that included discussion of Kurds either in the main title/headline or the lead paragraph. For instance, an article titled “Barzani is in Ankara” does not explicitly mention the word Kurd, but since Barzani is the leader of Iraq Kurdistan Regional Management, it was assumed that the article may have associations, and so it was treated as relevant. Further, the lead paragraph was checked if there were any references or associations of Kurds. For instance, if the Kurds are
portrayed positively or negatively, or if the Kurds are discussed in relation to the
democratization package (education in the mother tongue), a terrorist organization such as PKK/KCK, or willing to establish an autonomous state, then that news article was
considered relevant. Applying these rules, the census of relevant articles in the time period in these news outlets was 1,080 articles in Hürriyet and 1,032 articles in Sabah, a total of 2,112 newspaper articles. Within the census, the sample has been gathered by applying a systematic random sampling method. Every third article among 2,112 was chosen, and so
3 On the Hürriyet archives, the word “Kürt” has been searched. However, search results unavoidably
presented also results for the word “kurt” which means “wolf” in Turkish. Moreover, the word “kurt” is a common surname in Turkey. For that regard, it was crucial to distinguish the relevant content.
704 in total were analysed. 360 articles (51.1%) from Hürriyet (N= 360) and 344 articles (48.9%) from Sabah newspapers (N= 344) were manually coded.
Measures
The human coding process, in which the data was classified, covered steps of categorizing how Kurds were portrayed either in a positive or in a negative way. Due to the sensitivity of the issue, it was important to prevent researcher bias. For that purpose, only the manifest content of written communication was analysed (Riffe, Lacy & Fico, 2005). This, in the end, ensures the face validity of this analysis.
The codebook was developed based on an in-depth reading of a subsample of articles, as well as previous literature relevant to these discussion (See Somer 2005b). The codebook was composed of three sections and included twenty four items in total. The first group of items was about the fundamental details about the article itself, for instance, the date and length of the article. The second part included items to detect the content of the article through the variables that aim to measure how Kurds are referred to in the news article. Based on Murat Somer’s variables, Kurdish identity variables were created. The first variable asks if the article explicitly refers to someone Kurdish with the use of the word “Kurdish”. The second variable asks whether the article employs the term Kurd in reference to a group such as Kurdish party, Kurdish rebels. The third, to a cultural concept such as Kurdish language or Kurdish music; and lastly, to a place such as a Kurdish village and/or Kurdistan. This item is important to comprehend the discussions of Kurds’ requesting an independent/autonomous state within the borders of today’s Turkey. This item also provides information about whether the Kurds are discussed in relation to Iraqi Kurdistan, emphasizing the relations between Turkey and Iraq, and even Syria to some extent. These items were coded for their absence/presence in the article. When they were present, they were coded “1” and when they were absent, they were coded “0”.
The next set of variables aimed to measure if denigratory terms were present in the article. In other words, it measures whether Kurds are portrayed positively or negatively. For
that regard, it is very important to analyse the words selected to portray Kurds. The portrayal was first analyzed at the word level. For instance, it was checked if a positive or negative adjective used before the term Kurd. Positive adjectives were very often “(Muslim) brothers” and/or “loyal (to homeland)”. On the other hand, negative adjectives were usually “traitors”, “separatists” and/or “terrorists”.
Secondly, the portrayal was measured on the sentence level. Although the meaning of the sentence would mean the opposite, the appearance of any positive or negative identification of Kurds is regarded as a crucial tool for the portrayal of Kurds. Atteveldt, Kleinnijenhuis and Ruigrok (2008) explain this as “the power of words undergirds the persuasive appeal of political speeches, negotiation moves, eyewitness reports, and media coverage” (p.428). For instance, a soldier’s father says “We are Kurdish but we are not terrorists”. Despite the sentence implying that this Kurdish person is not a terrorist, the co-occurrence of certain words in media texts can create an implicit association in readers’ minds. To put it differently, through employment of particular word associations, certain issues can be primed in readers’ minds. This has been widely discussed, especially in marketing research: audiences/readers respond easier when two categories/words are associated/presented together than when they are not (Grimes & Kitchen, 2007). For that purpose, it is argued that some word associations can help readers shaping their semantic maps. Therefore, when words “Kurd” and “terrorist” appeared together in the same
sentence, this sort of quotations were coded as negative portrayal of Kurds. Lastly, for the overall use of these positive/negative adjectives per article were coded, in which “0” denotes the text used more negative words, “1” denotes positive and negative adjectives were fairly evenly mixed, “2” denotes the text employed more positive words. In case of absence of any positive/negative adjectives, it was coded “3”.
The third and last set of variables in the coding scheme measure any positive or negative associations made in the news article in reference to Kurdish. For this study, Mithun’s definition of ethnic identity is used. Mithun (as cited in Maslak, 2008, p. 88) describes ethnic identity as a concept that could derive “from common religious, physical,
linguistic, aesthetic or historical origins”. Based on this definition and Somer’s arguments of security/non-security associations of Kurds, association items were created. Each
newspaper article was coded in relation to the absence/presence of positive/negative associations. When an association was found, it was coded “1”; when not found, it was coded “0”.
For the overall use of positive/negative associations, the coding was done on a “1” to “5” scale in which “1“ denotes the articles used only negative associations, and “5” denotes the article used only positive associations. It was coded “2” when articles used more negative associations than positive ones (after counting of associations made), and it was coded “4” for the opposite situation. When there is a balance in the usage of these associations (i.e. if one PKK and one EU association co-exist in the article), it was coded as “3” which denotes neutral. Any discussion associating Kurds to a security issue is regarded as a negative association whereas any non-security discussion is regarded as a positive association.
In intricate situations, to exemplify, when it appears that a positive adjective was employed in a negative meaning, such as “We are not Muslim brothers”, for the existence of “brother” adjective, it receives “1” for the positive adjective (only if there are no other adjectives, the overall use of positive/negative adjectives can be coded “only positive”) and yet the meaning suggests the opposite. There was made an association of religion, and thus in the association items, it receives “1” for the presence of association and depending on other associations made in the article, the overall use of associations is coded.
Having finalized the coding, an intra-coder reliability test was conducted with 10% of the original sample. After coding 70 randomly chosen news articles, reliability of each item was measured by employing Krippendorff’s Alpha test. The alpha values ranged between .25 and .92, yielded satisfactory results for the majority of the variables (see Table 1 below). Attempts to include items that scored below the .67 threshold, as commonly accepted (Neuendorf, 2002), proved to be unsuccessful and therefore these items were
excluded from further analysis. For variables that scored below .80, only tentative conclusions can be made, as recommended by Neuendorf (2002). 4
Table 1
The Krippendorff’s Alpha scores of variables used in the analyses
Variable Name Krippendorff’s Alpha
Kurdish Identity - as a group .83
Kurdish Identity – as a concept .93
Kurdish Identity – as a place .83
Overall use of Positive/Negative Adjectives .77 Source referenced in the article .69
Kurdish Ethnicity .68
PKK association .83
Syrian/Iraqi War .80
Overall use of Positive/Negative Associations .67
Note. Only reliable items are shown. (See Appendix Table 2 for the scores of all items and notes.)
Results
4
Earlier, an inter-coder reliability test was conducted with a native-Turkish speaker who was trained for two days by practicing the coding of articles different from the originals used in the analysis. Later, 10% of the original sample was coded by two coders independently. Based on 70 randomly chosen news articles from the original coding, reliability of each item was measured by employing Krippendorff’s Alpha test. However, the majority of the items in the codebook yielded scores below the commonly accepted .67 threshold. (See Appendix, Table 1). These low scores can be an outcome of a poor description of items in the codebook or unsatisfactory/insufficient training of the coder. However, one should not disregard the complex and controversial nature of the topic. Due to these problems and time constraints, further analyses rely on intra-coder reliability.
Before revealing the main findings of the analyses, it is important to give background information about the data. A first interesting finding is that the year 2012 stands out among other years with the largest density of news articles covering Kurds (See Figure 1 in
Appendix). This could be an outcome of various real world developments – taken into account for further analyses – that potentially might prompt shifts in the media coverage of Kurds. Some examples include former Prime Minister Erdogan meeting an independent Kurdish representative for the first time, the kidnapping of a journalist and a representative from the Republican People’s Party by PKK militants, and some attacks by the PKK – which was organically tied to Kurds.
The second interesting outcome that the data provided is about the source of the article (see Figure 2 in Appendix). For instance, a majority of the news articles in Hürriyet were received from press agencies, namely from Anadolu Agency, a state-run press agency, and from Doğan News Agency, owned by the Doğan Media Group, (N=191). This suggests that although it is privately owned, Hürriyet’s inclusion of both state and business discourses may result in a rather “balanced” coverage of Kurds. On the other hand, many news articles in Sabah were journalist source (N=268), could possibly imply that the state discourse was the predominant one among others.
Another interesting finding concerns how Kurdishness is referred to in the news articles. The items created based on Somer’s items (except the “person” item, which yielded to be an unreliable one) revealed that Kurdishness can imply a combination of three major elements – belonging to a “group”, speaking the same “language”, inside a drawn border in which one can call a “state/nation”. This can add up to belong to a certain ethnicity, as the theoreatical arguments also put forward. Turkish news coverage, when referring to Kurds, mostly emphasized these three elements – which are supposedly creating a Kurdish ethnic identity. A principal component analysis with Varimax rotation on identity items indicated that Kurdish-group and Kurdish-place items can create a scale, yet it had a very low reliability score, Cronbach’s alpha= .21. Thus, the attempts to create a Kurdish identity scale were not successful.
One last critical information about the data is that the 21.7% of articles, the majority, published in Hürriyet newspaper gave voice to Kurdish sources (parties, leaders, members and representatives of formerly DTP – Democratic Society Party – and BDP –Peace and Democracy Party – as well as today’s HDP – People’s Democratic Party). The second highest share was given to the government voices with 20.6%. On the other hand, in Sabah, 18.3% of the news, the majority of articles, refer to the governing Justice and Development Party leader, members and representatives. The second most referred source is Kurdish people who do not belong to any political party (17.4%). This simple descriptive information can be sufficient to reveal many factual details about the newspapers’ ideological standings, supporting theoretical arguments claiming that Sabah is the voice for the government and Kurdish sources (See Figure 3 in Appendix).
To have more statistical ground in the analyses, the data is more delved into. Both dependent variables – the overall use of positive/negative adjectives and the overall use of positive/negative associations – were treated as interval scale variables to conduct further analyses. In other words, the difference between negative and neutral is regarded as the same with the difference between positive and neutral. Eventually, an independent t-test was conducted to compare if there is a significant difference between Hürriyet and Sabah with regards to first, their overall use of positive/negative adjectives and second, the overall use of positive/negative associations while discussing Kurds. Including the assumption of homogeneity of variances that was tested via Levene’s F test – overall use of
positive/negative adjectives: F(404)= 2.33, p= .128 and overall use of positive/negative associations: F(702)= 1.66, p= .198. There was no significant difference found between Hürriyet and Sabah newspapers for their scores of overall use of positive/negative
adjectives: t(404)= -1.13, SE= .08, p(two-tailed)= .260, 95% CI [-.26, .07] and overall use of positive/negative associations of Kurds t(702)= -.11, SE= .09, p(two-tailed)= .915, 95% CI [-.18, .16].
Mean scores of overall use of positive/negative adjectives and overall use of positive/negative associations in in Hürriyet and Sabah newspapers
Newspapers Overall Use of Positive (2)/ Negative (0) Adjectives
Overall Use of Positive (5)/ Negative (1) Associations
N M SD N M SD
Hürriyet 227 .54 .82 360 3.51 1.10
Sabah 179 .63 .85 344 3.52 1.17
It has been found out that both Sabah and Hürriyet applied negative adjectives referring to Kurds. Digging into the context of the news articles, it has been found that sources referred were addressing Kurds as a “problem” (Kurdish problem). Likewise, the overall use of positive/negative associations in the news articles imply that both newspapers used rather neutral/slightly more positive associations while referring to Kurds (see Table 2 above). In the light of this statistical information, both sub-hypotheses could not find a statistical support. Therefore, it is possible to claim that Hürriyet and Sabah newspapers do not differ in their overall use of positive/negative adjectives referring to Kurds nor in their employment of positive/negative associations while discussing Kurds in the news articles. Overall, it cannot be further discussed that the news coverage of Kurds in Hürriyet is more negative than Sabah newspaper; thereupon the first hypothesis is rejected.
Despite the fact that it was not possible to find quantitative grounds for the hypotheses, some quotes from the two different news outlets could possibly explain why both newspapers were not proposing a different coverage of Kurds and eventually why not any support for the hypotheses was found. To illustrate, news articles published on two random dates were chosen. The content of the news articles was identical, which facilitates making comparisons to find out if any potential discursive differences exist. Lead paragraphs
from two newspapers on June 1, 2011 are shown below. Hürriyet newspaper gives voice to a Kurdish source with no direct quotation:
Hürriyet Newspaper, 01.06.2011:
“Independent representative Leyla Zana from Diyarbakır, who is supported by Peace and Democracy Party, told that in case she enters the parliament again after the elections on June 12, she would wear a headscarf.”
As is seen from the quotation above, the Hürriyet newspaper stresses the importance of secularism by taking a rather defensive-nationalist standing. A Kurdish representative taking part in decision-making, and hew wearing headscarf in the parliament, are treated as threat elements. This implies that Kurds are discussed from a security-perspective despite there seemingly being a religious association with reference to the headscarf. On the same day, Sabah newspaper takes another person in its agenda, giving voice to the leader of the Republican People’s Party (CHP) by directly quoting him:
Sabah Newspaper, 01.06.2011:
“The CHP leader, who says ‘it is your brother Kemal’s sacred task to solve the Kurdish problem’, claimed “when we said ‘Autonomy’, they said ‘They are dismantling the country’. ‘The biggest separatist is Erdogan’.
This above quotation supports that Kurdish is usually negatively labelled with “problem”, an issue that requires a solution. Moreover, it is discussed together with Kurds’ request of an independent/autonomous state and how that was received by the governing party leader (Erdogan). That request seemed to be considered as “separatism”. Here Kurds are again discussed from a security perspective. These below quotations, different from the previous illustration, refer to the same sources, cover the same topics, yet they reveal it slightly different.
“Ahmet Türk, Independent Representative from Mardin, mentioned that besides being an education, culture, identity language; the Kurdish language should be an official language. He told: ‘All these have to be evaluated with one common politics. Kurds are now a power and their success came with alliance. Kurdish language raised many litterateur. Mehmet Uzun wrote a novel in Kurdish’. Over what Vice Prime Minister said ‘Kurdish is not a language of civilization’, Türk said: ‘The Kurdish language was used as an education language for years. It raised many litterateurs such as Ehmede Hani and Fegiya Tevra’.”
Sabah Newspaper, 03.03.2012:
“Ahmet Türk, who mentions that language is related with identity, culture and politics, told that ‘if language disappears many problems in other areas that support unity will be experienced. We request to have education in Kurdish language, using Kurdish language in politics and that it should be accepted as an official language. Vice Prime President says Kurdish language is not a language of civilization and it will not be used in education. Kurdish language has been a civilization and education language for centuries’ he said.”
Having exemplified possible qualitative explanations for the first hypothesis, it is possible to explain the analysis for testing the second hypothesis, in several steps. First, with the years as the independent variable, and the overall use of positive/negative adjectives and associations as dependent variables, a multivariate analysis of variance (MANOVA) test was conducted for both newspapers. The test results indicated that years significantly changed the overall use of positive/negative adjectives and associations in both
newspapers. Based on this statistical information, it cannot be argued that Hürriyet stays constant over time. Both newspapers indicate changes over the years of 2010-2014:
Hürriyet: F (8, 442)= 2.97, p= .003, Wilks’s Λ= .90, partial η2= .05; Sabah: F (8, 346)= 3.35, p < .001, Wilks’s Λ= .86, partial η2= .07. However, digging more into the data, it has been
found that this is only significant for the overall use of positive/negative associations. In other words, there is a significant effect of years only on the overall use of positive/negative
associations, both in Hürriyet and Sabah (See Figure 1 below). Thus, it can be proposed that the second hypothesis finds partial support.
Figure 1
Overall Use of Positive/Negative Asssociations in the News Articles, in Hürriyet and Sabah through 2010 -2014
Possible messages to be inferred from the above graph is that Sabah newspaper has significantly decreased the amount of negative-toned articles referring to Kurds, shown by the amount of “only/mostly negative” articles, and they have almost disappeared in 2013 and 2014. Likewise, Hürriyet also stopped publishing “only/mostly negative” articles in the same time periods. These findings eventually indicate that the overall use of positive/negative associations is usually independent from the political orientation of the newspapers but it could be related to external factors. These external factors could be that the sources referred in the news stories varied as well as the content of the news stories changed due to the real world developments with regards to any events/incidents involving Kurds.
Having demonstrated these results, it is possible to focus on possible real world developments in pre-defined time sequences, which could possibly prompt shifts in the news
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2010 2011 2012 2013 2014 2010 2011 2012 2013 2014 Hürriyet Sabah
coverage of Kurds in both newspapers. As discussed before, due to their political character, it is expected that Hürriyet will tend to stay rather constant while Sabah will tend to change its coverage over time. For that purpose, these news outlets were compared before and after conflictual and non-conflictual times, in relation to their overall use of positive/negative
associations. For non-conflictual time, the general elections on June 12, 2011 were chosen, and for conflictual time, the Gezi Park protests in May 28, 2013 were chosen. Despite election times could possibly create a tension between Turkish and Kurdish, it is not expected to be as tense as during the street demonstrations which resulted in civilian deaths. One potential pitfall, also of content analysis, is that both periods can include a combination of tension and non-tension periods.
To be able to make statistical inferences, six months - one month is considered as thirty days - , before and after tension and non-tension times were compared. An
independent t-test was conducted with General Elections in 2011 (before and after) as independent variable and the overall use of positive/negative associations as the dependent variable. It yielded a significiant result only for Hürriyet newspaper F(83)= .067, p= .800, t(83)= 3.40, SE = .23, p(two-tailed)< .001, 95% CI [.32, 1.23], while it produced a non-significant outcome for Sabah newspaper F(62)= .015, p= .900, t(62)= -.86, SE = .34, p(two-tailed)= .394, 95% CI [-.98, .39]. These results indicate that the overall use of
positive/negative adjectives does not change, from positive to negative or vice versa, in Sabah newspaper, but it does change in Hürriyet newspaper before and after General elections in 2011, which was defined as a non-tension period. In contradiction to what was expected in the second hypothesis, Sabah newspaper stays constant before and after conflictual times concerning overall use of positive/negative associations while discussing Kurds.
Table 3
Mean scores of overall use of positive/negative associations, before and after General Elections in 2011 and Gezi Park Protests in 2013
Newspapers Time General Elections in 2011 Gezi Park Protest in 2013 N M SD N M SD Hürriyet Before 41 4.07 1.03 71 3.59 .11 After 44 3.30 1.07 55 3.84 .14 Sabah Before 29 3.45 1.35 71 3.65 .12 After 35 3.74 1.38 45 3.95 .13
Having found earlier that the number of only/negative articles about Kurds decreased in 2013 in both newspapers, the year 2013 was taken into analysis. The end of May 2013 was marked with one of the largest mass civil movement in Turkish political history. Large street environmental protests, against the decision of the Municipality of Istanbul about demolishing Gezi Park in the centre of İstanbul’s Taksim Square and using the space for constructing a new shopping mall, received great attention, mostly from international media. Attention was strong due to the ever rising tension between security officers and civilians, in some cases resulting in civilian deaths (Yardımcı-Geyikçi, 2014). The results of an
independent t-test entail that the overall use of positive/negative associations both in Hürriyet: F(124)= 1.20, p= .276, t(124)= -1.39, SE = .18, p(two-tailed)= .166, 95% CI [-.59, .10], and Sabah newspaper F(114)= 1.89, p= .172, t(114)= -1.72, SE = .18, p(two-tailed)= .088, 95% CI [-.66, .05] has not changed before and after the protests. Although this was one of the biggest social unrest events in Turkish political history, ideologically different newspapers did not reveal their true political character in their coverage during this period. One explanation for this could be the strict state censorship on Turkish media during the protests (Yardımcı-Geyikçi, 2014).
Conclusion and Discussion
Comparing two opinion-leading yet ideologically different newspapers, this study expected to find differences between the Hürriyet and Sabah newspapers with regards to their news coverage of Kurds. This was analysed in relation to their usage of
positive/negative adjectives and associations while discussing Kurds. One major finding of the analysis indicated that the political ideology of the newspapers did not really matter in their portrayal of Kurds. In other words, Hürriyet and Sabah are not essentially different in their news coverage of Kurds, despite their divergent political character. Although they both tend to portray Kurds using negative adjectives – sources often call the Kurdish issue a “problem” that needs to be “solved – the overall use of positive/negative associations were revealed to be between neutral and more positive in both newspapers. This was
demonstrated with examples using negative labels that had a positive meaning at the
sentence level, such as “Not all Kurds are terrorists”. For this kind of quotation, the presence of a negative adjective was coded, whereas the meaning of the sentence suggested the opposite. When no negative associations follow, the overall use of positive/negative associations tends to be rather neutral or more positive when other associations are taken into account.
Although being a defensive-nationalist newspaper, and thus expected to have a rather negative discussions of Kurds, Hürriyet now seems to represent a rather positive state discourse to a certain extent. In contrast to what Sarıgil (2009) suggests, Hürriyet seems to leave its oppositional standing behind and tends to support Kurdish ethno-nationalist movements and groups. One can argue that this is an improvement in the discourse about the Kurds, but alternatively this can also imply that Hürriyet shifts from its political left standing towards the centre.
This similarity between Hürriyet and Sabah can be explained by the sources from which the newspapers retrieve their news. Although known to be an anti-government
newspaper, Hürriyet slotted a large space for news from a state-run press agency. Although including different voices in discussions can be interpreted as a sign for objectivity, the idea of Hürriyet referring to governing voices in a large share of news articles discussing Kurds seems to align with what McQuail (2010) put forward: that the ruling elite controls the information to legitimize their discourse.
Another major finding of the study indicated that Hürriyet changed its portrayal of Kurds over time, which was not expected. Similarly, Sabah showed changes over time as well – but this was expected. Changes for both newspapers only applied to their use of positive/negative associations while discussing Kurds, not to their use of positive/negative adjectives. The first reason both newspapers change could be that actually these two newspapers are similar in their portrayal of Kurds, as shown by the first hypothesis being rejected. An alternative explanation for the changes in both newspapers could be related to an accumulation of external factors.
The impact of two potential external factors was further analysed by comparing the news coverage of Kurds around two major events: the General Elections in 2011, as a non-tension event, and the Gezi Protests in 2013, as a non-tension event. Although the year 2012 had the largest density of articles referring to Kurds due to various events/incidents, the years 2011 and 2013 were more suitable in providing options for non-tension/tension periods. It was expected that Hürriyet would stay constant before and after tension and non-tension times, but the General Elections were found to affect the news coverage of Kurds. Hürriyet had rather neutral associations of Kurds after the General Election, while having positive associations before. This was unexpected – it could potentially be explained by other events taking place in the same time period. This underlines one of the pitfalls of content analysis: other events/incidents that may have happened in the chosen time periods could affect the outcome. Moreover, Sabah was expected to change, but this did not find statistical support. This presumably is an outcome of the stable agenda of the government towards Kurds.
Overall, it can be argued that, independent from their political alignments, two newspapers employ similar frames and thus contain commonalities in their news coverage of Kurds. This might indicate an alignment with what Kinder and Sanders (1990) argue: frames are embedded in political discourse. As Burlet and Reid (1998) claim, the discourse used in ethnic and/or racial discussions is “political”. Thus, this outcome may indicate a growing improvement in the discourse about the Kurds, considering the word “Kurd” was a
taboo in the 1980s (Romano, 2002; Somer, 2002). This eventually can explain the decrease in the application of a denigratory discourse over the time period analysed. This supports Somer’s (2005a) argument that new discursive strategies were developed by the Turkish state and media actors as a consequence of historical changes in the Turkish political landscape. To put it more specifically, the neo-Ottoman approach of the Justice and Development Party in which Kurds were aimed to be united with Turks under the umbrella term “ümmet” – by emphasizing the importance of religion as a uniting cement – seems to endure dominance in the Turkish media discourse. Consequently, the lines between left-wing and right-left-wing newspapers blur over time due to newspapers converging towards each other in the way they portray Kurds. Another explanation could be, as framing theories put, that communicators present information that resonates with their audiences’ existing schemas (Shoemaker & Reese, 1996). This could be regarded as an indication of changing audience profiles, which future studies can also include in their argumentation in more detail.
The major strength of this study is its choice of method: content analysis is the most appropriate method to analyse and compare large amounts of media content over time. Although this study endeavours to be one of the first systematic analyses in the literature by conducting a quantitative content analysis, the controversial nature of the topic of the study – the Kurdish issue is regarded as a sensitive subject (Somer, 2002; Romano, 2002) – makes it problematic to some extent.
Because Hürriyet and Sabah are highly circulated newspapers in Turkey, the
findings could presumably be generalizable to other mainstream-like newspapers. However, this can be tested in future studies with the inclusion of additional left-wing and right-wing newspapers. Moreover, the possible generalization is restricted to the chosen time-periods. Future studies can conduct a more in-depth analysis of events by further looking into events that could possibly create a sense of ethnic consciousness and perhaps including more recent events.
One limitation of this analysis is that it focuses on a specific case—that of Kurds in Turkish news media. However, in contemporary world politics, especially with the
establishment of nation-states, the recognition and representation of cultural and ethnic minorities becomes increasingly important. In certain cases, existing ethnic struggles in multi-ethnic countries have become tense. Therefore, this study provides a great potential not only for Communication Science, but also for Political and Social Science scholars, by being an exemplar analysis to investigate the portrayal of other ethnic groups in similar situations in other countries.
Another limitation is the exclusion of possible economic associations (e.g. economic underdevelopment of the South-East of Turkey, which may provide implications for the portrayal of the Kurds living in the region) due to the political scope of the study. Moreover, many items in the coding scheme proved to have low Krippendorff’s alpha values, which restricted further analysis and evaluations. Future studies are encouraged to employ quantitative analysis, enriching it by improving the socio-political variables and perhaps including novel variables such as the economic aspects of the issue.
Overall, this study may have repercussions for Turkey’s future ethno-politics as well as its neighbouring policies with Iraq and Syria, where a large of number of Kurds live in a semi-autonomous Kurdistan state. These can be interpreted as good news for the future of Kurdish-Turkish relations. The balanced/more positive portrayal of Kurds in the media could imply that the line between left-wing and right-wing newspapers is becoming less visible with the continuous liberalization of the discourse toward Kurds. This liberalization can be an outcome of the rising consciousness and politicization of the Kurdish identity (Somer, 2002, 2004; Romano, 2002). This presumably was triggered by the EU’s promotion of
democratization and human rights. This enabled Turkey to take a step in improving the conditions of Kurds – starting by giving cultural rights – with the implementation of the Democratic Package, which eventually can facilitate to cease the ongoing armed-conflict between Kurds and Turks. This paves a way for scholars to further delve into this issue.
References
Al, S. (2015). Elite discourses, nationalism and moderation: a dialectical analysis of Turkish and Kurdish nationalisms. Ethnopolitics: Formerly Global Review of Ethnopolitics, 14(1), 94-112.
Barışı Getireceğim. (2011, June 1). Sabah Newspaper. Retrieved from http://www.sabah.com.tr
BDP’li vekillerden marş şov. (2012, March 3). Sabah Newspaper. Retrieved from http://www.sabah.com.tr
Berger, A.A. (2007). Media and society: a critical perspective. Available from https://books.google.nl/books/about/Media_and_Society.html.
Burlet, S. & Reid, H. (1998). A gendered uprising: Political representation and minority ethnic communities'. Ethnic and Racial Studies, 21(2), 270-287.
Circulation numbers. Retrieved from http://www.medyaradar.com/tirajlar.
Çakmur,B. & Kaya, E. (2010). Politics and the mass media in Turkey. Turkish Studies. 11(4), 521-537.
Çeliker, A.G. (2009). Construction of the Kurdish self in Turkey through humorous popular culture. Journal of Intercultural Studies, 30(1), 89-105.
Demir, I. & Zeydanlioğlu, W. (2010). On the representation of ‘others’ at Europe's borders: The case of Iraqi Kurds. Journal of Contemporary European Studies, 18(1), 7-23. De Swert, K. (2012). Calculating inter-coder reliability in media content analysis using
Krippendorff’s alpha. Retrieved from http://www.polcomm.org/wp-content/uploads/ICR01022012.pdf
De Vreese, C.H. & Schuck, A.R.T. (2006). Between risk and opportunity, news framing and its effects on public support for EU enlargement. European Journal of
Communication, 21(1), 5–32.
Entman, R. (2007). Framing bias: Media in the distribution of power. Journal of Communication, 57, 163–173.
Entman, R.M. & Gross, K.A. (2008). Race to judgment: Stereotyping media and criminal defendants. Law and Contemporary Problems, 71(4), 93-133.
Ergin, M. (2014). The racialization of Kurdish identity in Turkey. Ethnic and Racial Studies, 37(2), 322-341.
Goffman, E (1974). Frame Analysis: An essay on the organization of experience. Retrieved from
http://is.muni.cz/el/1423/podzim2013/SOC571E/um/E.Goffman-FrameAnalysis.pdf.
Grimes, A., & Kitchen, P.J. (2007). Researching mere exposure effects to advertising. Theoretical foundations and methodological implications. International Journal of Market Research, 49(2), 191-219.
Hallin, D.C. & Mancini, P. (2004). Comparing media systems. Three models of media and politics. Cambridge: Cambridge University Press.
Iskander, E. (2012). The ‘mediation’ of Muslim–Christian relations in Egypt: the Strategies and discourses of the official Egyptian press during Mubarak’s presidency. Islam and Christian–Muslim Relations, 23(1), 31–44.
Jonkman, J. & Verhoeven, P. (2013). From risk to safety: Implicit frames of third-party airport risk in Dutch quality newspapers between 1992 and 2009. Safety Science, 58, 1–10. Kinder, D, R. & Sanders, L.M. (1990). Mimicking political debate with survey questions: The
case of white opinion on affirmative action for blacks. Social Cognition 8, 1, 73–103. Kürt marşıyla Kürt dili konferansı. (2012, March 3). Hürriyet Newspaper. Retrieved from
http://www.hurriyet.com.tr.
Maslak, M.A. (2008) School as a site of Tibetan ethnic identity construction in India? Results from a content analysis of textbooks and Delphi study of teachers' perceptions. Educational Review, 60(1), 85-106.
McQuail, D. (2010). McQuail’s Mass Communication Theory (6th edition). Sage Publication.
Mutlu, S. (1996). Ethnic Kurds in Turkey: A demographic study. International Journal of Middle East Studies, 28(4), 517-541.
Neuendorf, K.A. (2002). The content analysis guidebook. Thousand Oaks, CA: Sage Publications.
Panayırcı, U.C. & Iseri, E. (2014). A content analysis of the AKP's “honorable” foreign policy discourse: The nexus of domestic–international Politics. Turkish Studies, 15(1), 62-80.
Riffe, D., Lacy, S. & Fico, F.G. (2005). Analyzing media messages. Using quantitative content analysis in research (2nd edition). New York NY: Routledge.
Romano, D. (2002). Modern communications technology in ethnic nationalist hands: The case of the Kurds. Canadian Journal of Political Science / Revue Canadienne de Science Politique, 35(1),127-149.
Sarigil, Z. (2010). Curbing Kurdish ethno-nationalism in Turkey: an empirical assessment of pro-Islamic and socio-economic approaches. Ethnic and Racial Studies, 33(3), 533-553.
Scheufele, D. A. & Tewksbury, D. (2007). Framing, agenda setting, priming: The evolution of three media effects models. Journal of Communication. 57, 9-20.
Shoemaker, P. J. & Reese, S.D. (1996). Mediating the message. Theories of influences on mass media content (2nd edition). USA: Longman Publishers.
Sezgin, D. & Wall, M.A. (2005). Constructing the Kurds in the Turkish press: a case study of Hurriyet newspaper. Media, Culture & Society, 27, 787–798.
Simon, A.F. & Jerit, J. (2007). Toward a theory relating political discourse, media, and public opinion. Journal of Communication, 57, 254–271.
Somer, M. (2002). Ethnic Kurds, Endogenous identities, and Turkey's democratization and integration with Europe. Global Review of Ethnopolitics, 1(4), 74-93.
Somer, M. (2004).Turkey's Kurdish conflict: Changing context, and domestic and regional implications. Middle East Journal, 58(2), 235-253.
Somer, M. (2005a). Defensive- vs. liberal-nationalist perspectives on diversity and the Kurdish conflict: Europeanization, the internal debate, and Türkiyelilik. New Perspectives on Turkey, 32, 73-91.
Somer, M. (2005b). Resurgence and remaking of identity: Civil beliefs, domestic and external dynamics, and the Turkish mainstream discourse on Kurds. Comparative Political Studies, 38(6), 591-622.
Somer, M. & Liaras, E.G. (2010). Turkey’s new Kurdish Opening: Religious versus secular values. Middle East Policy, 17(2), 152-165.
Van Atteveldt, W., Kleinnijenhuis, J. & Ruigrok, N. (2008). Parsing, semantic networks, and political authority using syntactic analysis to extract semantic relations from Dutch Newspaper articles. Political Analysis, 16, 428–446.
Van Bruinessen, M. (1998) Shifting national and ethnic identities: the Kurds in Turkey and the European Diaspora. Journal of Muslim Minority Affairs, 18(1), 39-52.
Van Gorp, B. (2005). Where is the frame? Victims and intruders in Belgian press coverage of the asylum issue. European Journal of Communication, 20(4), 484–507.
Yardımcı-Geyikçi, Ş. Gezi Park protests in Turkey. A party politics view. The Political Quarterly, 85(4), 445-453.
Yeğen, M. (2007). Turkish nationalism and the Kurdish question. Ethnic and Racial Studies, 30(1), 119-151.
Zana: Meclise türban takıp girebilirim. (2011, June 1). Hürriyet Newspaper. Retrieved from http://www.hurriyet.com.tr.
Appendix
1. Kurdish Study Codebook
Codebook – Content Analysis: News Coverage of the Kurdish
The codebook is composed of three parts. In the first part, the coder(s) will code fundamental details about the article itself. This includes the name of the newspaper where the article comes from, the date, the news outlet that published the story, the main title of the article (only), and how long the news story is.
