Brussels via Athens:
How much is Macedonian attitude on the EU contingent on perception of Greek politics
Ivo Bosilkov
Student ID: 10831630
Master’s Thesis
Graduate School of Communication
Master’s programme Erasmus Mundus Journalism, Media and Globalisation
Supervisor: Rachid Azrout
Abstract
Can the attitude towards a single EU member state act as a proxy for support of EU integration among citizens of an EU candidate state? Public opinion research has identified a range of determinants of support for the integration process, concluding that the influence of media is often overlooked among them. In this experimental study (n = 181), Macedonians have been exposed to different news frames regarding neighbouring EU member Greece, which has been blocking Macedonia’s EU accession due to a bilateral dispute. The aim was to discover how different combinations of frame valence (positive and negative) and frame issue arising from two of the strongest EU support determinants (economic and national identity) would impact their attitudes on EU integration and levels of Euroskepticism. Results have shown that individuals in fact exhibit less support for the EU when exposed to a positive frame about Greece, indicating to a possible case of reverse mobilization, while they also consider the economic factor to be more prominent than the threat of national identity loss, despite the identity-centred nature of the dispute with Greece. However, the perception of national identity as important to EU accession does increase after exposure to an identity frame, and it is also found that political knowledge moderates Euroskepticism. Due to the complex relations of the two countries and the salience of Greece-related frames in Macedonia, experimental research about framing effects and EU attitudes would require exceptional subtlety.
Keywords: EU, public opinion, Macedonia, Greece, framing, experiment, national identity, economy, media, Euroskepticism
Introduction (I) A non-EU approach to Euroskepticism
Although public support for EU integration in the Republic of Macedonia has been overwhelmingly high since the 1990s, the frustrations from a stagnating process, coupled with other international developments such as the economic crisis in Europe have resulted with a steady decrease of the integration support in the recent years. A survey done by the International Republican Institute (IRI) has shown that the percentage of respondents supporting Macedonia membership in the EU has dropped from 96% in September 2009 to 84% in November 2012, while opposition has grown from 4 to 12 percent in the same period.
This fall is far from dramatic, but it is certainly indicative of a more general tendency that has swept not only Macedonia, but EU members themselves. The more frequent
manifestations of Euroskepticism in Macedonian public discourse are echoing sentiments about the project across the continent, with contemporary research even recognising that
Euroskepticism has shaken the confidence in further European integration and forced to redefine theories about it (Hooghe & Marks, 2007). In an effort to understand this phenomenon, social scholars who have been investigating it have found that its roots exist in factors such as discontent about the economic performance of the EU, the identity and cultural threat of
integration, institutional distrust and elite influence. With the studies on Euroskepticism being a part of a more wider study on EU public support aiming to identify determinants of attitude towards the EU (Taggart & Szczerbiak, 2002; Kopecky & Mudde, 2002), it is noticeable that the
core components of Euroskepticism are a reflection of the broader theoretical framework of EU public support.
In this regard, the academic community is in general agreement over a wide spectre of specific models that explain citizens’ opinion about the EU, through the economic (or utilitarian) perspective, cultural (or value/identity) perspective, and the political (democracy, partisanship) perspective. Without getting deeper in these theories at this point, it is sufficient to say that they have been successfully applied both in the context of the West (Maier & Rittberger, 2008; Hooghe & Marks, 2005) and the more recently integrated Eastern European countries (Cichowski, 2000; Ehin, 2001), but much less in the context of EU candidate countries. As crucial decisions about the future of the EU, such as enlargement and the Constitutional Treaty have been put to the fate of national referendums, we are witnessing a significant shift of European integration from an elite-dominated process to one dependent on public attitudes. For this reason, making sense of public opinion is increasingly important in social science.
(II) The missing link: media effects
While the models explaining public opinion are essentially covering the summary of concepts that drive EU attitudes in a variety of settings, one aspect has been considered relatively overlooked - the one of media effects. As most citizens don’t directly experience and participate in high level EU politics, they rely on information from media coverage to form their opinions, with media giving cues to citizens on how to position themselves on issues concerning European integration. Indeed, as Azrout et al. (2012) conclude, media are an important factor of
understanding dynamics in support for EU enlargement, and accordingly, media are likely to determine levels of Euroskepticism among EU citizens (de Vreese, 2007).
However, as demonstrated by these examples, prior research on media and public attitude towards the specific topic of EU enlargement mostly done in the Western academic community has predictably focused on how citizens of countries already in the EU perceive this process (even the studies conducted in the context of Eastern European countries such as
Cichowski and Ehin reveal patterns of EU attitude in countries already in the Union). With regards to media effects, a range of studies on this topic (Lecheler & de Vreese, 2010; Maier & Rittberger, 2008; De Vreese et al., 2012; Schuck & De Vreese, 2006) has focused on the concept of news frames and the effects of exposure to them, confirming it as a reliable, effective and influential way in which media compel people to think about EU issues.
As opposed to what makes EU citizens think in certain ways about the admission of candidate countries, this study is seeking to answer what it is that influences Macedonian citizens’ attitude towards EU integration, and whether certain frames push it towards
Euroskepticism. Officially a candidate state for EU membership since 2005, the Republic of Macedonia has witnessed its approximation towards the strategic goal of European integration increasingly stalled in the last decade. While several factors contribute to this - some internal, like the Macedonian national-populist policies since coming to power of Nikola Gruevski’s government (Ivanovski, 2013) and others external, such as the infamous enlargement fatigue in the EU (Szolucha, 2010) - in the most simplified political context it ultimately boils down to one single reason. The main obstacle for Macedonia's entrance in the EU is considered to be the
ongoing profoundly identity-centered dispute about the country's name with neighbouring Greece (Pop-Angelov, 2010; Koneska, 2014).
This study aims to continue the path of previous research of media effects on public opinion, by showing the perspective of a country that is aspiring for EU membership,
furthermore locked in a unique stalemate with an EU member. Hence, it can potentially provide novel insight contributing to existing theoretical predictors in this domain. In order to do so, it intends to explore how differing domestic media news frames on Greece impact the public opinion on the EU and the integration process. The study will adapt the experimental designs and use the theoretical concepts introduced by some of the aforementioned studies. The following section will include comprehensive theory review (including background of the dispute with Greece) and according hypotheses. That will be followed by a methodology section, after which the results of the empirical study will be presented and discussed.
Theoretical review (I) How framing works
In recent times the concept of framing has received a considerable amount of attention in political communication. Frames are defined as 'patterns of interpretation that are used to classify information and process it efficiently' (Lecheler & de Vreese, 2010, p. 77). It is through frames that citizens make sense of politics, since the way an issue is characterized in news reports can have an influence on how it is understood by the audience (Scheufele & Tewksbury,
2007). Chong and Druckman (2007, p. 100) make a crucial distinction about the term, stating that an ‘individual frame’, which refers to an individual’s cognitive understanding of a given situation, is different from a ‘media frame’, which refers to the words, images, and presentation style that, for example, a media outlet uses when relaying information to the audience. These two conceptualizations of frames take place in different processes, systematized by Scheufele (1999) - the media frame corresponds to the ‘frame building’ stage, while the individual frame is related to the process of ‘frame setting’ and ‘individual level effects’. It is in the interaction of these processes that we can discover what kind of impact frames have on attitudes and whether they are effective. Although Chong and Druckman (2007) are specifically interested in the outcomes of competing frames on people’s attitudes, in their typology of framing research designs they declare the asymmetric one-sided design as the most common one, where the change of individual opinion through exposure to a single frame is measured. Analyzing the effects of isolated frames is therefore a stepping stone towards attempting to make sense of a diverse frame environment.
Several scholars apply this type of design in experimental studies, but additional
conceptual clarification of framing is required in order to achieve accuracy in the research goals. According to Entman (1993), the frame determines whether most people notice and how they understand and remember a problem, as well as how they evaluate and choose to act upon it. Entman suggests that frames promote a 'moral evaluation' (p. 52). This implies that frames have inherent valence by suggesting positive or negative aspects, solutions or treatments, and given this valence, news frames can be expected to influence public support for various policy measures. Therefore, by valence it is suggested that 'some frames are indicative of good or bad
and (implicitly) carry positive and/or negative messages (De Vreese & Boomgaarden, 2003, p. 363). This is the kind of approach also used by Schuck and de Vreese (2006) in their study of public support for EU enlargement, expressing valence in terms of risk (negative) and
opportunity (positive), thus showing that individuals exposed to the opportunity frame show significantly higher levels of support than their ‘risk’ counterparts. Lecheler and de Vreese (2010) arrive to the same conclusion - positive frames increase support - in this case in a slightly altered context of public support for the integration of one specific candidate country, namely Serbia. De Vreese et al. (2012) focus on the case of Turkey, as the most visible and controversial among EU citizens, their findings regarding exposure to valenced frames in line with the other studies. However, they exhibit an additional breakthrough by empirically proving an assumption with background in existing literature; namely, that exposure to negatively valenced frames will have stronger impact on attitude than exposure to positively valenced frames.
Concluding the theoretical breakdown of framing effects with two important notes; firstly, Sniderman and Theriault (2004) demonstrate, context in the shape of prior values of individuals subjected to frames matters greatly. This is corroborated by Scheufele (2000), who claims that models of framing effects have to pay more attention to the individual orientations and attitudes of media users that exist prior to exposure to certain news frames. Following this rationale, if individuals are exposed to a media frame that invokes their belonging to a specific group (an identification that would naturally imply a distinctive set of values), it can be expected that their attitude will be more affected than if this collective identity isn’t mobilized. If the group in question is Macedonians, it can be assumed that due to a negatively predisposed
cognitive association with neighbouring country Greece, especially in the context of EU integration, the frame will be made increasingly salient.
Second, and in relation to this, a number of studies include political knowledge as a key moderator in processing of frames. In their experiment on support for EU enlargement after exposure to risk and opportunity frames, Schuck and de Vreese (2006) show that individuals with low political knowledge are more strongly affected by news frames. However, as they point out, while political knowledge is influential in frame analyses, it yields inconsistent results. Kinder and Sanders (1990) results back their findings, while Nelson et al. (1997) have found the opposite. Again, it would be important to reevaluate these findings from the perspective of an EU-candidate country, for which existing theory is scarce. Despite the fact that literature is divided as to whether more politically knowledgeable individuals are more susceptible to
framing effects than less knowledgeable individuals, in line with the findings of Lecheler and De Vreese (2010) about attitudes on Serbia's EU integration, I predict that those with low knowledge will be more affected by the frames, because ‘they do not possess the elaborate mental
“stockpile” to fight off framed information (p. 79). In the given context, the expectation is that the those who are incapable of understanding the frames will be extremely few, therefore the term ‘low knowledgeable’ is applied to a group relative to the one of the ‘highly knowledgeable’, who have the ability to counter-frame and are therefore more resistant. This distinction will be operationalized in the measurement of the knowledge moderator variable.
(II) Macedonian perception of Greece: why Euroskepticism?
The obvious question arising from the above is why would Macedonians feel negatively towards Greece in the context of EU integration? The answer lies in the complexity of historical developments in the region extending all the way back to ancient times up until modern day Republic of Macedonia and its bid for EU membership. Since Macedonia’s declaration of independence from the dissolving Yugoslav federation in 1991, Greece has ferociously objected to its use of the term ‘Macedonia’ as the name of the new state, arguing that it implies
irredentism through territorial aspirations and appropriation of Hellenic cultural heritage (Pop-Angelov, 2010, p. 4). The Macedonian question on the Balkans has been a matter of controversy since almost two centuries ago, with the withdrawal of the Ottoman Empire, the emergence of nationalism and newly formed nation-states with ambitions to control the territory of
geographical region called Macedonia and assimilate its ethnically diverse population. As this region was partitioned between Greece, Bulgaria and Serbia after the Balkan wars, its Slavic inhabitants were denied an opportunity to consolidate a distinct national identity, until the Second World War, when political circumstances in the shape of communist victory in the war forced Serbia to renounce its part, allowing the creation of a constitutive Macedonian republic in the new socialist federation of Yugoslavia (Floudas, 2002).
While the issue laid dormant due to geopolitical considerations until the Yugoslav breakup, Macedonia and its people, the ethnic Macedonians, faced a fierce opposition to its international recognition and its very existence by the Greek state, immediately after announcing independence. Arguing that the name Macedonia is an exclusively Greek cultural property,
through the continuity of Hellenism since ancient times and the age of Alexander the Great, Greece has inflicted measures to suppress the recognition of the new state, even delaying its admission in the United Nations until 1995, when the Interim Accord was signed allowing entrance under the provisional name ‘Former Yugoslav Republic of Macedonia’ (Pop-Angelov, 2009).
Since then, Greece has continued to contest Macedonian national identity and the country’s right to use its constitutional name in international organizations, culminating with the veto for NATO membership at the summit in Bucharest in 2008 (International Crisis Group, 2009). As Euro-atlantic integration is Macedonia’s strategic goal, the prospect for EU
membership has been simultaneously blocked, despite official candidate status since 2005 and multiple recommendations from the European Commission to open accession talks. At the EU summit in June 2008, Athens has succeeded in making the resolution of the name issue an additional condition for further progress in EU integration (International Crisis Group, 2009). The new conditionality was reiterated on 8 December by EU foreign ministers, who declared that it is essential for Macedonia ‘to maintain good neighbourly relations, including coming to a mutually acceptable solution to the name issue’. Even the French foreign minister was
unequivocal: ‘Without the resolution of the name issue, we cannot move forward, because this issue has to be solved first’.
It is fairly evident that the conflict over the name with Greece is the fundamental hurdle for Macedonia’s EU accession. One study shows that Macedonians overwhelmingly support integration in the EU with 90.9% of the respondents answering affirmative to the question
(Markovic et al., 2011, p. 10), however when conditioned by name change, the numbers are opposite (Gallup, 2009, p. 7; Klekovski, 2014, p. 20). Due to the nature of EU decision making, where the accession of new members is accepted through unanimous vote among the existing members, the EU stands in solidarity with Greece over the name dispute. Because of all this, it can be argued that Macedonians’ sentiments toward the EU are contingent on their feelings toward Greece, as a looming entity that is inseparable from Macedonia’s prospect of EU integration. The fact that Greece is considered the regional economic and political giant and its the longest standing EU member is an additional reason why the EU may be evaluated through the perception of Greece. This kind of reasoning is based on Anderson’s (1998) argument that in the absence of direct knowledge, public opinion about the EU is created through proxies. While Anderson claims that these proxies could be system support (satisfaction for the performance of institutions), attitudes towards the incumbent national governments, or simply political parties’ attitude towards the EU, the unique position of Macedonia could imply another type of proxy; that is, the attitude towards an individual EU country.
In short, as a member state of the EU (and one that in recent times gained even more salience in the context of recent EU economic developments), Greece and its politics towards Macedonia, for the most part vis-a-vis the name dispute, is considered to be often cognitively and symbolically associated with EU politics itself. With the general outlook being that Greece's politics is the main culprit behind Macedonia's lack of progress in the integration process, hostility towards Athens is largely harbored. A valuable finding would be confirmation that this hostility is projected towards Brussels as well (since the two political entities are perceived as
inseparable in the integration decision making process), and whether this antagonism is manifesting through deeper Euroskepticism.
It is important to emphasize that Euroskepticism itself is a concept that is complex and not fully synonymous with 'support for the EU’. As McLaren (2004) asserts, people can be indifferent about the EU, or even fearful that it may pose a threat to their identity or culture, but still support the country’s membership due to economic and political benefits. In the broadest sense, Euroskepticism carries the meaning of doubt and distrust on the subject of European integration, with the degree of this distrust ranging from moderate 'European integration has gone as far as it should go' to extreme 'outright rejection of membership in the EU’ (Flood, 2002, p. 73). However, it must be noted that again in these theoretical conceptualizations, it is generally the ‘within-EU’ attitudes, from a Western point of view that are observed, with a possibility that opposition to further enlargement can be interpreted as Euroskepticism. As most of the Balkan countries have generally favourable opinion for joining the EU, in their case Euroskepticism should be narrowed down to a more subtle concept, where citizens are opposed to the EU through different motivations.
This way not only the confusion with the broader concept of ‘EU attitude’ is avoided, but also more accurate picture of the levels of Euroskepticism is achieved. Taggart and
Szczerbiak (2002) make the distinction between between soft Euroskepticism, which is not principled, rather representing concern about certain EU policies, and hard Euroskepticism, where downright opposition to the whole project of European integration is expressed as a principle. Conti’s (2003) reverse description makes it even clearer: ‘functional Europeanism’,
where there is no principled support for European integration, but rather pragmatic approval of the EU, and ‘identity Europeanism’, for a principled commitment to European integration akin to what was once called ‘federalism’. These categorizations suggest a need to operationalize
Euroskepticism on a scale level in the methodological section, but the main objective is to assess Euroskepticism as a broader principle on a more ideological level.
(III) Determinants of public support for the EU
A crucial semantic, conceptual and methodological distinction arises from the fact that Macedonians' support for the country's EU accession shouldn't necessarily correspond with their opinion about the European project in general, as the considerations driving these positions might be of different nature. This distinction is perhaps related not to the valence of the frames, but their type. The previously mentioned experiment by De Vreese et al. (2012) on the effects of news framing on support for Turkish membership in the EU has employed an economic frame, a geopolitical security frame and a cultural frame, all valenced either positively or negatively, showing that an economic threat frame impacts support more negatively than the cultural or security threat frame. Similarly, Maier and Rittberger (2008) have used articles about economy, democracy and culture, with the results indicating that the democracy frame creates the biggest change in subjects’ opinion. For the construction of their experiment, de Vreese et al. (2012) follow the theoretical distinction between generic frames which are typically applied across a range of topics, and issue-specific frames, which they utilize, describing them as intrinsically related to certain topics or news events. This dimension is also applicable for the case of testing
the effect of frames on Macedonians, as the name dispute is a specific topic that certainly doesn’t really transcend thematic boundaries.
The choice for the particular set of determinants is based on a vast amount of literature on public support for the EU. Karp and Bowler (2006) consider three main indicators for support of specifically EU enlargement: the feeling of European identity, instrumental self-interest and relationship with EU institutions. They find the self-interest to be the strongest indicator, and many others agree that it is calculations of economic cost-benefit that determines support for the EU (Gabel, 1998). However, a distinction needs to be made between egocentric and sociotropic utilitarianism (McLaren, 2004). While the egocentric reflects Gabel’s (1998) conceptualization, where individuals would support the EU based on human capital, or their personal perceptions of gain and loss, the sociotropic goes beyond economic benefits expected by an individual and instead focuses on the perception of national economies; as Kentmen (2008, p. 491) puts it ‘the citizens hold the EU accountable for national economic conditions’. These two economic parameters correspond with what Ehin (2001) calls micro and macro-economic models of support, and they represent the utilitarian perspective on EU attitudes.
On the other hand, another group of scholars believes that while utilitarian
considerations are important, it is identity that overshadows them as the most crucial factor of EU support. Hooghe and Marks (2005) say that although there is a possibility for developing complementary European identity, national identities are much more deeply rooted in people’s minds than a distant idea of European integration, and hence the thought of the national
identity. This idea draws from social identity theory, where the in-group sees the out-group with a negative bias, and European integration erodes these traditional norms of ‘us’ and ‘them’ of European nation states, creating the sense of threat for the nation. In a previous study, Hooghe and Marks (2004) say that ‘emotional’ group loyalties and attachments about the national are formed earlier than the ‘rational’ attitudes towards a supranational political entity, and this may create a conflict when there is a ‘reallocation of authority across levels of government’, among those who have a stronger national attachment or exclusive national identity. They find that relative to economic considerations, it is these conceptions of group membership that are more powerful.
The fear of loss of national identity is instrumental to this argument, as it manifests itself through antipathy and even hostility towards other nations, or even entities that are supranational, but are threatening national integrity, such as the European Union in this case. According to Carey (2002), national identity can be conceptualized in three ways, through intensity of feelings toward one’s country, through attachment to territorial entities and perhaps most indicatively, through the perceived threat from other cultures. The rationale of the third conceptualization is that people who are fearful of the process or European integration leading to their language being used less, or their national identity and culture becoming less distinct, are expected to hold a more negative view of the EU. As McLaren (2002) further argues, the EU is likely to be seen as having homogenizing effect on the member states, threatening the uniqueness of national cultures. If an analogy is made with the utilitarian considerations, this fear would correspond with the sociotropic dimension, since again in evaluating public policies, it is societal
level needs rather than own personal needs that have prominence. In a nutshell, individuals who have a stronger national attachment are more likely to have negative attitude about the EU.
There is no reason either why the identity dilemma wouldn’t apply for a EU applicant country. Still, it would be relevant for the sake of illustration to weigh the economic and identity determinants in these settings. Unlike EU-based studies, research conducted in Turkey, surveying the opinion about potential EU membership show that economic considerations edge out identity concerns (Kentmen, 2008; Cartoglu & Kentmen, 2011). However, Turkey’s situation is also different from Macedonia in a number of ways. For example, the attachment to Islam introduces a whole new variable to cultural identity, but such that is inapplicable to the Macedonian case, one which is unprecedented from a completely different perspective, by the contention of the country’s name and national identity. Apart from religion, as part of cultural identity (De Vreese et al., 2012), the impact of other determinants is explored as well as operationalized in several studies; such as the idea that anti-immigrant sentiments are antecedent for EU enlargement support through the process of ‘othering’ (Azrout et al., 2012), or similarly that attitude towards minorities is an indicator (McLaren, 2002). Another line of research points out that
post-materialistic values in individuals is related to higher support (Inglehart, 1977), others that the inclination for democracy and its inherent values like stability of institutions, rule of law and human rights is increasing EU affection (Maier and Rittberger, 2008; Ehin, 2008; Cartoglu and Kentmen, 2011).
Nevertheless, because in the particular case of Macedonia it is primarily the economic and identity evaluations that receive increased salience through the proxy of Greece, I
incorporate them as focus of the research. While McLaren (2004) concedes that there is no solid evidence that threat to national identity has a stronger impact than the utilitarian evaluations about EU attitudes, Hooghe and Marks (2004) claim the opposite, emphasizing however that this varies from one country to another depending on the political context. It is because of the
idiosyncratic circumstances of the Macedonia where identity is contested, that it can be assumed that identity concerns are going to depress EU enthusiasm even more than economic concerns.
(IV) Hypotheses
With the theoretical frame established, the hypotheses can now be posited. Based on parts (I) and (II), I formulate the following hypotheses:
H1a - Exposure to negative news frames about Greece will result with lower levels of support for EU accession of Macedonia.
H1b - The impact of exposure to negative frames is stronger than the impact of exposure to positive frames.
H1c - Levels of support for Macedonia's accession in the EU in individuals with more political knowledge will be less affected by the news frames than the individuals with less knowledge.
Because of the conceptualization of Euroskepticism as separate from attitude about EU accession, I formulate the following hypothesis as second:
H2 - Negative news frames about Greece will result with higher levels of hard Euroskepticism.
Based on part (III), where it was argued about the comparative strength of economic and identity frames vis-a-vis attitudes about the EU, I posit the third hypothesis:
H3a - The impact of identity frames in relation to Greece is bigger than the impact of economic frames concerning Greece in terms of the support for EU accession of Macedonia.
And finally, in relation to the previous hypothesis, the last one is derived from the connection between the established determinants of public support with framing theory, to make a full circle for the theoretical review. As a group of scholars suggests that framing effects are mediated by belief importance (Nelson et al., 1997), meaning that framing affects individuals by altering the perceived importance of some aspects of an issue, according to De Vreese et al. (2012), this perspective suggests that trade-offs are made between different beliefs or
considerations when opinions are expressed. Consistent with the previously laid out theory on public support, an assumption can be made that subjects exposed to utilitarian frames will see economy as more important, with the same effect present in the identity frame.
Methodology
Aiming to evaluate the impact of different news frames about Greece and Greek politics in relation to the attitude about EU integration and potential Euroskepticism, the study draws inspiration from existing research of media effects and public opinion, specifically where news frames are presented to subjects with expectations of according attitude change. The research design of this study is experimental, as it uses a manipulation for the independent variable (news frame) for various groups in order to investigate the effects on the dependent variable, giving control over content and exposure and thus permitting to draw conclusions about causality. Consequently, the study is quantitative and deductive in its nature, but also exploratory (since it attempts to establish a causal connection in an insufficiently researched setting).
The choice of experiment as the method in this case is naturally the most suitable one, because it reduces confounding variables to a minimum, therefore minimizing systematic errors and increasing internal validity (Azevedo et al., 2011, p. 233). The study employs a 2x2 design, combining the two issue frames (identity and economic frame), with two valences (positive, negative), to generate four experimental conditions in which subjects are randomly divided into, in addition to a control group. The experiment does not include a pre-test, but only post-test to make a between-subject comparison across conditions. The total number of subjects that participated in the experimental survey was 183, meaning approximately 35 individuals per experimental condition - generally considered a sufficient number to observe behavioral patterns.
The population is ethnic Macedonians; in this sense it is important to emphasize the rationale that no members of other ethnic groups (such as Albanians that are citizens of the
Republic of Macedonia) are included in the experiment, since the cultural/identity frame would only be effective vis-a-vis the dispute about the country's name with Greece and the contested national identity of ethnic Macedonians, rendering intuitive expectations that other ethnic groups' attitude change would be insignificant in terms of the research question. This decision is justified by a recent survey released by the Macedonian Secretariat for European Affairs, showing that 82.1 per cent of ethnic Macedonians said that for them the preservation of the name 'Republic of Macedonia' was more important than the EU accession of their country, while in contrast, 77.8 percent of ethnic Albanian respondents opted for EU and NATO membership before the name (Marusic, 2010). The support of EU accession in Macedonia underscores this assumption; ethnic Macedonians’ support for EU accession has fallen to 54% in 2014, while in the ethnic Albanian minority it remains as high as 91%, hence the cumulatively high score on a country level in support for EU accession in other surveys (Marichikj & Petkovski, 2014, p. 28). Indeed this distinction between Macedonia’s ethnic groups is essential for the proper utilization of national identity as a core concept behind this study.
Ideally, probability sampling would render best results due to attaining representative sample, however circumstances allowed to conduct only convenience sampling. Nevertheless, the sampling procedure attempted to achieve at least minimal representativeness of the
population in terms of demographics (age, education and political orientation) by additionally looking into performing quota sampling during data collection in order to identify
underrepresented subgroups and recruit their members. The recruiting was done through
Facebook, also involving an element of snowball sampling as subjects were asked to forward the Qualtrics survey to others using the same recruiting method. After being recruited and accepting
to participate, subjects were randomly assigned between the experimental groups within the distributed survey.
The subjects were given material that begins with a demographics survey regarding their age, education, partisanship and general ideological orientation, followed by a political knowledge test, after which they were exposed to the stimulus material and finally asked to answer the short survey. Analysis whether the experimental conditions are correlated to these pre-test variables showed a generally even distribution of subjects within the experimental conditions per demographic categories, with groups (despite efforts to achieve a representative sample) all consisting of individuals predominantly of younger age, high education (Bachelor), leftist political orientation and unaffiliated to a political party. (See Appendix A)
Subjects were informed that they are filling a questionnaire about support for EU integration, but the implication of Greece as an independent variable was not revealed until the debriefing upon completion of the survey. The political knowledge test determining the
moderator variable consisted of five factual closed-ended questions, three related to Macedonian politics and two about the EU, ranked by difficulty: 'Which two main parties form the
Macedonian government?', 'Who is Macedonia's minister of foreign affairs?', 'Which party does Radmila Sekerinska belong to?', 'Is Croatia a member of the EU?' and 'Who is the president of the European commission?', with an index scale from 0 to 1 with higher values indicating higher level of knowledge, later rendering two groups of high and low knowledgeable subjects. The 1
choice of a mix between national and EU questions to assess political knowledge is justified by
It is worth noting that the phrasing of the first question in the process of translation to Macedonian may have 1
slightly affected the answers, as part of the subjects didn’t interpret the question as which parties form the government now, rather which parties form the government in general as election winners.
the fact that national knowledge has been used as indicator in EU related studies (Hobolt, 2007), and solely EU-related knowledge questions in general often yield low threshold means (Schuck & De Vreese, 2006).
The positive and negative valence of the news frames as the independent variables were defined by inherent attributes, that is 'by employing language that classically indicates a positive or negative issue development' (Lecheler & de Vreese, 2010, p. 80). This explains how the valence was constructed in the articles, with consideration of the tone and expert quotes determining the positive or negative frame. The economic and identity issue frames will emphasize the economic progress (or failures) of Greece and identity tolerance (or rejection). The articles were similar in size, structure and wording, ensuring uniformity in the interpretation patterns, and therefore high reliability. All of them referred to a fictional session of the Council of the EU, where the first and last chapters were the same, describing more general activity at the session, while the middle two paragraphs are manipulated for the purpose of the framing, backed up with expert quotes from real people and referring to real media for the impression of
authenticity (See Appendix B).
The questionnaire was constructed so the concepts defined in the theoretical part are operationalized in a way which would make answering the hypotheses possible, hence assuring high internal validity. In the first part, three questions are asked intending to evaluate attitude towards Greece as a mediating variable, through a feeling thermometer, followed by two items designed to determine the belief importance with regards to the utilitarian and the identity frame after being exposed to the stimulus. Items about EU attitude and Euroskepticism as dependent
variables were closed-ended statements with a choice on a 5-point Likert scale, ranging from fully agree to fully disagree, producing interval data scales. Items about EU attitudes were phrased to reflect support for membership, such as a hypothetical vote in a referendum about joining EU and importance of EU membership for Macedonia in the long term future. To complement them, attitudes about economy and national identity as concepts related to the EU have been operationalized through items surveying the level of perceived benefit from EU membership and perceived job creating potential from membership (in the case of economy), or fear of losing national identity/language and perceived compatibility with EU cultural values (in the case of national identity). In the last part, to determine more detailed Euroskepticism levels, I adapted Kopecky and Mudde (2002) concept of the four ideal types of orientations towards the EU (Eurorejects, Europragmatics, Eurocritics and Euroenthusiasts) into items regarding general support for EU integration as opposed to Macedonia’s accession; the items were also done on a 5-point scale to ensure compatibility, pinpointing the degree to which subjects agree or disagree with the ideal of EU integration, its current model and the required power of EU institutions.
Generating scales
Before running the tests the survey items were combined to generate the scales for the said variables, after performing factor and reliability analyses. These tests have shown that the three items describing EU attitude have good internal consistency (α = .81), as do the three utilitarian items (α = .82), the three forming the scale about attitudes on Greece (α=.63), and Euroskepticism (α=.65). The principal content analysis is also showing similar values in the
component matrix in each of the scales, except for the Euroskepticism variable, where the item ‘the current model of the EU is a good approach for European integration’ has a .47 score on the component matrix, unlike the higher values of the other three items. For this reason, it was possible to make a reliable scale about the attitudes towards EU integration, the attitudes about the economy vis-a-vis EU integration and overall feelings about Greece, but not about the general Euroskepticism, since the mentioned item was correlating poorly with other three items, which together give a better perspective on Euroskepticism.
Regarding the attitude on national identity and the EU, one item in the cluster was not compatible with the rest, therefore having to be removed from the scale to assure consistency. In this case, one component (measuring perception of shared values between Macedonia and the EU) was the only one that had eigenvalue over one (1.70), scoring .23 on the component matrix, compared to above .90 in the other two items. If the item was included in the scale, Cronbach Alpha = .57, while without it, it rises to .80. For this reason it was also considered a variable to be measured separately and not included in the scale that explains fears about national identity loss in the case of EU accession. This scale was converted so that higher mean scores reflect less fear from losing national identity and/or language loss.
Manipulation check
The manipulation check was done through three factual questions about the articles, requiring answers which were later coded ‘correct’ (1) and ‘incorrect’ (0). The first question referred to all four experimental conditions, asking what is the event that the text describes,
giving the options “NATO summit”, “EU parliament session”, the correct one “EU Council session”, and a “not sure” option. Only 53.6 percent of the subjects answered correctly, with 5 percent answering that it is about a NATO summit, 13.6 that it is about an EU parliament session and 27 percent not sure. The two other questions were supposed to measure if people perceived the valence of the frame correctly. The one given to subjects in the economic frame was stated “Politicians are speaking about Greece’s economic reforms” with the correct answers being positively for the positive frame and negative for the negative one. A total of 74.3 percent answered correctly to this question, with 85.2 percent in the negative economic condition, and 67.4 percent in the positive. The question given to the subjects in the identity frame was stated “Greek politicians say that Macedonia constitutional name can be used in”, with the options “all international institutions”, “some international institutions” and “no international institutions”. The total of correct answers is 81.4 percent, rendering manipulation successful, however in the negative condition the percentage of correct answers (no international instutions) is much higher (95.3%) than the correct answers (some international institutions) in the positive condition (59.3%).
Results
The hypotheses on dependent variables like EU attitudes, utilitarian/national identity attitudes and broader Euroskepticism were tested by comparing means between the experimental conditions. The first test was supposed to show whether subjects adjust their attitude towards Greece after being exposed to the manipulated frames. By means of one way analysis of variance
(ANOVA) it was found that individuals who were exposed to a negative frame, indeed have reported more negative feelings towards Greece (M = 58.87, SD = 9.47) than those who were exposed to a positive frame (M = 60.06, SD = 8.83), yet both conditions have more negative feelings than the control group (M = 61.15, SD = 9.50), F (2, 178) = 0.82, p = .442. Similarly, those who received an identity frame had more negative feelings (M = 58.88, SD = 9.15) than those who got an economic frame (M = 60.06, SD = 9.15), with the control again having most positive responses (M = 61.15, SD = 9.50), F (2, 178) = 0.82, p = .444. However these results should be taken with caution since they didn’t have statistical significance.
The ANOVA showing the main effect of valence of the manipulated frames to levels of support for EU accession has rather counterintuitively demonstrated that respondents exposed to negative news frames on Greece in fact had higher levels of support for EU accession (M = 4.12, SD= 0.98), compared to those that saw a positive frame (M = 3.88, SD= 1.04). This result
effectively rejects the first hypothesis H1a, as the findings seem to actually be reversed from those expected, F (2, 178) = 0.96, p = .385. However, as noted here, these results are not statistically significant, although there was a homogeneity of variance according to Levene’s F (2, 178) = 2.10, p = .125. These findings also render H1b incorrect, as the negative frame does yield a bigger support than the positive frame when compared to the control group (M = 3.98, SD = 1.11), but evidently the direction of the impact is reverse to the expected, and also the significance level doesn’t reach the necessary margin, so the impact can be due to chance.
Additionally, when a further analysis is conducted regarding the attitudes about the economic benefits from a potential EU accession, these contradictory findings are confirmed, as respondents faced with a negative frame about Greece’s economy see more benefit from EU accession (M = 3.67, SD = 1.03), than those faced with the positive (M = 3.53, SD = 0.91). Again, statistical significance indicates there is no significant difference between the groups, F (1, 176) = 1.23, p = .270. Statistical significance eludes the study in the case of fear of national identity loss too; although subjects in the negative identity condition fear losing identity and language use in case of EU accession (M = 3.64, SD = 1.12) more than those in the positive identity condition (M = 3.80 SD = 1.30), this difference is again not significant, F (1, 176) = 2.52, p = .616.
Similarly as with H1, the findings are running counter to H2 as well. Although the assumption was that negative framing of EU member Greece will reflect to the attitude about the Union itself, the ANOVA test has shown that subjects who were exposed to the negatively framed articles have scored higher on the Euroskepticism scale (M = 3.88, SD = .86), than those in the positive condition (M = 3.62, SD =.76). It is important to note that the bigger the mean, the less Euroskeptic actually people are. These findings however also lack statistical significance, F (2, 178) = 1.76, p = .175. Results are similarly counterintuitive when testing the item that didn’t fit in the scale; subjects in the negative condition believe that the current model of the EU is a good approach towards European integration (M = 3.06, SD = 1.01) more than those in the positive condition (M = 2.91, SD = 0.97) and the control group (M = 2.93, SD = 1.08). While this finding, like the previous one, also achieves homogeneity of variances, it also fails to achieve statistical significance, F (2, 178) = 0.40, p = .669.
Having set out to test Hooghe and Marks (2004) argument about the strength of national identity as a determinant for EU integration in comparison to economic evaluations, H3a has been proven to be incorrect, as the mean difference for support between the positive and negative economic frame (M difference= -0.43, p = .091) is greater than the mean difference between the
positive and negative national identity frame (M difference= -0.02, p = .937), F (4, 176) = 0.84, p
= .501, therefore the difference is not significant. In this case the reverse direction of frame influence is again visible, also in the case when analyzing difference between the control group, with both positive (M difference= 0.21, p = .360) and negative (M difference= -0.23, p = .382)
economic frame, compared to the difference with positive (M difference= -0.07, p = .800) and
negative (M difference= -0.09, p = .706) identity frame.
When checking to see if the prediction that individuals exposed to a frame that
highlights the economic or national identity issue will have their considerations adjusted to the issue frame they received, the findings show that subjects in the economic frame do believe the economic factor is more important than the national identity factor, F (2, 178) = .43, p = .649, and vice versa, with those exposed to the identity frame believing that identity is more important, F (2, 178) = 3.31, p = .039 (see Table 1). While the first finding falls short of the desired
statistical significance, therefore not confirming H3b, for the second p <0.05, which indicates that being exposed to a frame that emphasises national identity has increased the considerations of national identity as important regarding EU integration among subjects. The post-hoc test shows that while there is a significant difference between the identity condition and the control (M difference= 0.62, p = .034), there is no such difference between the identity and economic
Table 1. Mean differences for belief importance
In order to find whether political knowledge has a moderating effect on the framing effects, first the knowledge index was divided to create two groups of individuals with high and low knowledge. Consequently, an interaction model was used, using the attitudes on EU
integration scale as a dependent variable. The findings firstly demonstrate that subjects with more political knowledge have consistently higher scores for EU support in every combination of frame and valence, F (1, 171) = 9.37, p = .003. Furthermore, the mean differences are greater between subjects from the group of low knowledge across all categories, which is especially noticeable in the identity frame, where those with low knowledge have visibly lower support when exposed to the negative frame. Also, the highly knowledgeable subjects’ totals are similar between the economic and national identity frames, while the ones with low knowledge differ more in the mean scores (see Table 2). All this being said, H1c cannot be confirmed because the differences are not statistically significant, F (1, 171) = 0.1, p = .911 and also Levene’s F (9, 171) = 2.07, p = .035, indicating that the assumption of equal variances has been violated.
Issue frame Importance of economic factor Importance of identity factor
Economy 4.33 (0.97) 3.80 (1.25)
Identity 4.24 (1.08) 4.13 (1.25)
Control 4.15 (0.91) 3.51 (1.23)
Table 2. Mean differences for support of EU integration
Although not one of the hypotheses, this has prompted curiosity to explore whether knowledge has a moderating effect on other dependent variables. While there wasn’t any significant difference for attitudes about economy or national identity, there has been a statistically significant difference between groups with high and low knowledge exposed to a positive or negative frame valence, demonstrating that those subjects with low knowledge in the negative condition (M = 3.33, SD = 0.93) are more euroskeptical than those with the same level of knowledge in the positive condition (M = 3.48, SD = 0.84), F (1, 171) = 4.17, p = .043. However among those with high knowledge, the results are confirming the peculiar reverse tendency (Table 3).
Issue frame Valence Low knowledge High knowledge Overall
Economy positive 3.43 (1.00) 4.07 (0.99) 3.77 (1.04) negative 3.74 (0.99) 4.44 (0.84) 4.21 (0.94) total 3.53 (0.99) 4.23 (0.94) 3.94 (1.02) Identity positive 3.78 (1.23) 4.26 (0.88) 4.05 (1.06) negative 3.61 (1.18) 4.25 (0.92) 4.07 (1.02) total 3.69 (1.18) 4.25 (0.89) 4.06 (1.03) N 64 117 181
Table 3. Mean differences for Euroskepticism (higher values is less)
Discussion
The main rationale behind this study stems from an established deficiency of scientific research in the scholarship of public opinion on EU integration from two crucial aspects. First, the relative lack of attention towards media as a source of influence for the formation of attitudes on EU matters, mostly considered a secondary determinant, trailing utilitarian considerations, as well as the so-called “affective” considerations (Maier & Rittberger, 2008), which, among other issues such as political values or social perceptions, also include the fear of loss of the national to the supranational. The latter, along with the utilitarian factor were used as thematic concepts which were supposed to highlight the functioning of the media, as this study tried to focus on media content in order to address this deficiency. The second deficiency concerns the obvious unilateral flow of public opinion research on EU enlargement, where the attitudes of current EU nationals regarding expansion with candidate countries are being probed. Analyzing
Macedonians’ attitudes, as citizens of an EU applicant state, was meant to give an example of the
Frame valence Low knowledge High Knowledge Overall
Positive 3.48 (0.84) 3.74 (0.67) 3.62 (0.75)
Negative 3.33 (0.93) 4.11 (0.72) 3.88 (0.86)
Control 3.54 (0.99) 3.90 (0.84) 3.80 (0.89)
opposite directional dynamics regarding EU attitudes and reveal a perspective about EU enlargement outside of the Western-centric approach.
The effort to assume media reports on Greece and corresponding attitudes as a proxy (Anderson, 1998) for attitudes on EU integration proved to be a tricky one. Unlike De Vreese and Boomgaarden (2003), this study hasn’t found any significant changes in terms of policy support, which means that the valenced news frames used in the experimental conditions didn’t have any effect. Looking at previous research, this outcome is uncommon, because the essence of framing theory is that media frames influence publics in a variety of ways. There are obvious reasons why this could be the case, many of them sample-related. The number of 181
respondents is by all standards sufficient for an experiment of this type, however the
randomization procedure generated groups that are numerically uneven and perhaps too limited to establish a pattern when taking into account the lack of demographic variance. Other design-related anomalies that could explain the unstable results include the impersonal nature of the experiment’s online administration, possibly decreasing subjects’ attention and dedication, while the complex nature of the name dispute between Macedonia and Greece also renders
conventional theoretical concepts and measures of fear of national identity loss inadequate, creating another completely different pitfall.
Yet, all this is not to say that the study had no results, since if significance is set aside, a puzzling pattern of attitude change emerges. Individuals exposed to a positive frame are more euroskeptic and less supportive of EU accession. While baffling at first, this finding can be explained. First of all, its important to emphasize that support for EU accession is still the
predominant attitude among Macedonians, although it is steadily decreasing in the past years. It is therefore unlikely that subjects would exhibit a dramatic change in attitude from the standard level of EU support, or experience a sudden increase of Euroskepticism. As the dispute over the country’s name is the most salient and pertaining issue in Macedonia since its independence, not only as a matter of foreign policy, but all-pervasive even to micro-societal level, it is safe to assume that most people have an already developed stance towards EU accession in relation to the name dispute, one that would be difficult to upset through exposure to media content. Furthermore, the salience of the issue in Macedonian society makes it practically impossible for a single article on the topic to induce a massive opinion shift. As previous studies about framing effects that also used an experimental design have found, exposure to a single frame is an unreliable method of tracking individuals’ opinion, in a situation where they are constantly exposed to multiple news frames on a daily basis. This study simply doesn’t account for competitive frames.
Related to this, another potential explanation for the unexpected findings is the realism level of the manipulated stimuli. In the given context, where both the troubled economic
situation in Greece, and especially the developments over the name dispute are receiving
intensive and detailed media coverage in Macedonia, it shouldn’t come as a shock that a positive “spin” on either of these two news topics would be taken with a grain of salt. Several scholars (Iyengar, 1991; De Vreese & Boomgaarden, 2003) in their studies have used existing stories that have been already broadcasted, despite recognizing that it could jeopardize internal validity. The decision of increasing internal validity at the expense of external validity in this study has possibly had a negative effect on its overall output, as the intricacies regarding the name dispute
with Greece require a more subtle approach to manipulation. In this case, although the
manipulated material’s objective was to imitate actual media content as faithfully as possible, through attributing to the article similar structure and language as actual news, it was still not genuine media content and therefore scientifically imperfect. Indicative to this is the fact that manipulation check for both positively valenced conditions yielded significantly smaller success rate than for the negative frames. People can simply see through the fictional news stories because they are too familiar with the context to be duped.
In this sense, it would be prudent to establish the characteristics of stories in Macedonian media that can be associated with negative or positive valence, as well as the economic or identity issue frame, before plunging into frame manipulation which could result with a potential perception of implausibility. As noted repeatedly, the name dispute case and the relationship with Greece is exceptionally complex for many reasons; there isn’t a similar case among EU candidate states, which leaves little room for manoeuvre when designing the stimulus materials. The case is similar with Greece’s economic struggles which are difficult to be
portrayed in a positive manner, especially in the latest period which coincided with this study, when Greece is widely speculated to leave the Eurozone. However, if a more systematic analysis is conducted about actual news content regarding Greece in Macedonian media (especially with regard to EU accession) extrapolating from these findings, the valence and issue frame could be more accurately presented. NVO Infocentar (2013) investigates how Macedonian media covers Bulgaria, finding that ‘the bulk of media content presents Bulgaria in a negative view’ and that ‘there is the selective approach and the manner of presentation of information served to the citizens, used by the media to perpetuate and reproduce the discourse of conflict’. If a similar
analysis is performed about Greece, it would lead to more representative experimental frames and more genuine response, reducing the dilemma of how to present positive content to people who basically encounter only negative content in real life.
Still, if we take the results from this study at face value, their most theoretically reliable interpretation is via manifestation of the psychological concept of “motivated skepticism” (Ditto and Lopez, 1992) according to which people tend to scrutinize more thoroughly information that is inconsistent with a previous conclusion or outcome. Hence, if respondents receiving a positive frame show less support for EU accession, it could mean that they simply processed the
information in a way that allows them to argue with it and reach a more skeptical conclusion, closer to their previous position and opposite of the expected frame directed effect. This reverse effect has been demonstrated by Schuck and De Vreese (2008), in the context of mobilization in EU referendums, where people faced with positive news about EU referendums were more mobilized to vote against in them because they were motivated to prevent an outcome
incongruent to their prior beliefs. The same principle could apply in this study, explaining why individuals who were exposed to a positive frame were keen to oppose the newly acquired information that runs counter to the status quo of their beliefs. Furthermore, if motivated individuals perceived the positively valenced experimental frames as weak, then it is not only likely to reject them, but also to move away from the position they advocated (Chong & Druckman, 2007). Because of the salience of Greece in Macedonian public sphere, it is safe to assume that an overwhelming majority of the subjects would be motivated, and because it contradicts prior beliefs (in the case of the positive frames) and doesn’t resonate with consensus values, it is safe to assume that they don’t see the frame as strong.
Although lacking in statistical significance, findings also show that individuals exposed to economic frames have experienced stronger attitude change depending on frame valence than those exposed to a frame about national identity. Even though this contradicts Hooghe and Marks’ (2004) findings, it reflects the contemporary dynamics in EU public opinion, which are reported by more recent research as well: the notion that in the current context of the economic crisis and unstable European currency, the economic evaluations have once again increased their prominence. As de Vreese et al. (2012) state, such a situational factor can’t be neglected, and this wasn't the case for earlier studies because of the context which called for a more
post-materialistic approach. Regardless of the other findings, this study corroborates previous conclusions that political knowledge is an important confounding factor when it comes to framing effects (Lecheler & De Vreese, 2010). Just as in EU member states, in Macedonia as well people with low political knowledge are more likely to have more negative opinion on the EU, but as this study also shows, when Euroskepticism is observed the less knowledgeable are also more likely to be influenced by frames in the desired direction than the high knowledgeable, which are reverse mobilized. This is an important implication to practical matters, since the strength of media content and elite cues needs to be observed within the distinction between informed and uninformed citizens in society. For the future, it would be interesting to include other related concepts as moderators, such as political interest, emotions (Lecheler et al., 2013), or post-materialistic values (Inglehart, 1977).
Despite all the shortcomings this study has still provided data about the mechanisms that influence public opinion regarding EU attitudes. Indications that media play a big role in the collective opinion, and therefore the decision making processes in EU integration have been
reinforced; frame effects are functioning, for which evidence enough is the confirmation that individuals believe the issue frame they were exposed to is of greater importance, and the moderating effect of political knowledge. While the unique context between Macedonia and Greece apparently requires an approach that is more diligent and creative, perhaps suggesting that future empirical studies should be even more expansive in manipulating the range of emphasis frames, explanations do exist that are in line with previous research about certain patterns of behaviour when confronted with different frames. As national identity is an extremely entangled issue in the Macedonia-Greece-EU triangle, it is more difficult to infer conclusions that can be generalized about other EU candidate countries, but this very fact can also mean that there is a lot of room for exploration regarding these concepts, with potential benefits for
References
Anderson, C. (1998). When in Doubt, Use Proxies: Attitudes toward Domestic Politics and Support for European Integration. Comparative Political Studies, 31(5), 569-601.
Azevedo, L., Canário-Almeida, F., Almeida Fonseca, J., Costa-Pereira, A., Winck, J., & Hespanhol, V. (2011). How to write a scientific paper—Writing the methods section. Revista Portuguesa De Pneumologia, 17(5), 232-238.
Azrout, R., van Spanje, J., & De Vreese, C. (2012). When News Matters: Media Effects on Public Support for European Union Enlargement in 21 Countries. JCMS: Journal Of Common Market Studies, 50(5), 691-708.
Carey, S. (2002). Undivided Loyalties: Is National Identity an Obstacle to European Integration?.European Union Politics, 3(4), 387-413.
Çarkoğlu, A., & Kentmen, Ç. (2011). Diagnosing Trends and Determinants in Public Support for Turkey's EU Membership. South European Society And Politics, 16(3), 365-379.
Chong, D., & Druckman, J. (2007). A Theory of Framing and Opinion Formation in Competitive Elite Environments. Journal Of Communication, 57(1), 99-118.
Cichowski, R. (2000). Western Dreams, Eastern Realities: Support for the European Union in Central and Eastern Europe. Comparative Political Studies, 33(10), 1243-1278.
Conti, N. (2003). Party Attitudes to European Integration: A Longitudinal Analysis of the Italian Case. EPERN Working Paper No 13. Sussex: Sussex European Institute.
de Vreese, C. (2007). A Spiral of Euroscepticism: The Media's Fault?. Acta Politica, 42(2-3), 271-286.
de Vreese, C., van der Brug, W., & Hobolt, S. (2012). Turkey in the EU?: How cultural and economic frames affect support for Turkish accession. Comp Eur Polit, 10(2), 218-235.
Ditto, P., & Lopez, D. (1992). Motivated skepticism: Use of differential decision criteria for preferred and nonpreferred conclusions. Journal Of Personality And Social Psychology, 63(4), 568-584.
Ehin, P. (2001). Determinants of public support for EU membership: Data from the Baltic countries. Eur J Political Res, 40(1), 31-56.
Entman, R. (1993). Framing: Toward Clarification of a Fractured Paradigm. J Communication, 43(4), 51-58.
Flood, C. (2002). The Challenge of Euroscepticism. In J. Gower (Ed.) The European Union Handbook 2nd Edition. (pp 73-84). London: Fitzroy Dearborn Publishers.
Floudas, D.A. (2002). Pardon? A Conflict for a Name? FYROM’s Dispute With Greece Revisited. The New Balkans: Disintegration and Reconstruction, Columbia University Press, New York, 1-11.
Gabel, M. (1998). Public Support for European Integration: An Empirical Test of Five Theories. The Journal Of Politics, 60(02), 333-354.
Gallup Balkan Monitor (2009). Focus On: Perceptions of the EU in the Western Balkans. Insights and Perceptions: Voices of the Balkans. Retrieved from www.balkan-monitor.eu
Hobolt, S. (2007). Taking Cues on Europe? Voter competence and party endorsements in referendums on European integration. Eur J Political Res, 46(2), 151-182.
Hooghe, L. & Marks, G. (2004) Does Identity or Economic Rationality Drive Public Opinion on European Integration? PS: Political Science and Politics 37: 415–20.
Hooghe, L. & Marks, G. (2005). Calculation, Community and Cues: Public Opinion on European Integration. European Union Politics, 6(4), 419-443.
Hooghe, L., & Marks, G. (2007). Sources of Euroscepticism. Acta Politica, 42(2-3), 119-127.
Inglehart, R. (1977). The Silent Revolution. Princeton: Princeton University Press.
International Crisis Group (2009). Macedonia’s Name: Breaking the Deadlock. Europe Briefing 52. 1-14
Ivanovski, H. (2013). The Macedonia-Greece dispute/difference over the name issue: mitigating the inherently unsolvable. New Balkan Politics, 14, 48-80.
Iyengar, S. (1991). Is anyone responsible?. Chicago: University of Chicago Press.
Karp, J., & Bowler, S. (2006). Broadening and deepening or broadening versus deepening: The question of enlargement and Europe's 'hesitant Europeans'. Eur J Political Res, 45(3), 369-390.
Kentmen, C. (2008). Determinants of Support for EU Membership in Turkey: Islamic
Attachments, Utilitarian Considerations and National Identity. European Union Politics, 9(4), 487-510.
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.
Klekovski, S. (2014). Sporot za imeto 2013. Macedonian Centre for International Cooperation, 1-32.
Koneska, C. (2014). Policy Consensus During Institutional Change. In S. Keil & B. Stahl (Eds.), The Foreign Policies of Post-Yugoslav States: From Yugoslavia to Europe (pp. 97-121), Palsgrave MacMillan.
Kopecky, P., & Mudde, C. (2002). The Two Sides of Euroscepticism: Party Positions on European Integration in East Central Europe. European Union Politics, 3(3), 297-326.
Lecheler, S., & De Vreese, C. (2010). Framing Serbia: the effects of news framing on public support for EU enlargement. European Political Science Review, 2(01), 73.
Lecheler, S., Schuck, A., & de Vreese, C. (2013). Dealing with feelings: Positive and negative discrete emotions as mediators of news framing effects. Communications - The European Journal Of Communication Research, 38(2). 189-209.
Maier, J., & Rittberger, B. (2008). Shifting Europe's Boundaries: Mass Media, Public Opinion and the Enlargement of the EU. European Union Politics, 9(2), 243-267.
Markovic, N., Ilievski, Z., Damjanovski, I. & Bozinovski, V. (2011). The role of the European Union in democratic consolidation and ethnic conflict management in the Republic of Macedonia, Regional Research Promotion Programme, 1-26.
Marichikj, B. & Petkovski, L. (2014). The political culture, europeanization and fears in Macedonia. Macedonian Centre for European Training. 1-31. Retrieved from http:// mcet.org.mk/en/dokument.asp?cnd=98
Marusic, S. J. (2010, July 14). Macedonia Survey Raises Euroscepticism Concern. Balkan Insight. Retrieved from http://www.balkaninsight.com/en/article/macedonia-survey-raises-euroscepticism-concern
McLaren, L. (2002). Public Support for the European Union: Cost/Benefit Analysis or Perceived Cultural Threat?. Jop, 64(02). 551-566.
McLaren, L. (2004). Opposition to European integration and fear of loss of national identity: Debunking a basic assumption regarding hostility to the integration project. Eur J Political Res,43(6), 895-912.
McLaren, L. (2007). Explaining Opposition to Turkish Membership of the EU. European Union Politics, 8(2), 251-278.
Nelson, T., Oxley, Z., & Clawson, R. (1997). Toward a Psychology of Framing Effects, Political Behaviour 19(3), 221–246.
