Final An Any Om

A patchy web: The European network of  newsflows. A network analysis    Master Thesis 

This study investigates the structure of the European network of newsflows. According to recent literature in the field, the horizontal integration of national public spheres is a constitutive element of a European public sphere. Based on a sample of news items in 25 countries, we present a comprehensive depiction of the mutual coverage of member states in the EU by applying network analysis. The results show that the network of newsflows is shaped by the economical and political power of the involved countries as well as by their cultural and historic proximity. Moreover, we identify two structural gaps: Smaller countries are generally marginalized and there is a clear division into an Eastern and a Western European public sphere. The duration of the membership of a country in the EU, however, has a positive effect on its integration in the network. In contrast to other recent studies in the field we find little evidence of the existence of a coherent European public sphere.

European Newsflows

Introduction  The portrait of the European Union in the news media across the continent is a mosaic. It consists of the depiction of its institutions: the Parliament, the Commission, and the Council. But there is more to it. Europe is not only what happens in Brussels, but also what happens in Berlin, Riga, or Lisbon. Therefore, the mosaic also inherits the portrayal of its member states and their representatives on the European level. The depiction of Europe is also a mosaic in a different sense. The elements that form the mosaic are distributed across time and media. Everyday only a few events - if any - that concern the Union in the widest sense make the news in a specific newspaper of TV news broadcast. The coverage of the EU in one news outlet is only a snapshot that deviates significantly from the picture that is presented elsewhere. On a random day, we might not find a single news item concerning the EU in the largest newspaper of Lithuania, Lietuvos Rytas, whereas the lead story of the Dutch Volkskrant might be about the Italian president nominating female top models as prospective candidates for the next European parliamentary elections (Arends, 2009). Hence, the mosaic, that is the portrait of Europe, looks differently every day in every European news outlet. There are, however, a number of underlying patterns in the coverage of the European Union, especially with regard to the depiction of fellow member states in national news media. To give an example: The results of a general election in UK are usually discussed extensively all across Europe, whereas Estonian general elections are more or less neglected in the majority of other European countries. It is the aim of this study to identify these patterns of asymmetric coverage of European member states in the news media. This is done by integrating the depiction of the EU members presented in the news outlets of 25 European countries over the course of three weeks into one common picture by using network analysis. This novel analysis technique allows us to depict the network of newsflows in the European Union. When looking at the structure of the newsflows in the EU we can discover its gaps as well as its strongest ties. Are events in particular countries more likely to be picked up than events in other countries? Moreover, we propose an explanatory model of the identified structural features of the network of European newsflows. The model is based on news value theory as well as on the output of an extensive body of empirical studies in the field of foreign news coverage research. Which characteristics of a country can explain that it is noticed more often than 2

European Newsflows others? And does the volume of coverage only depend on characteristics of the country covered, or also on country attributes of the country in which the news are being published or broadcast? Until today, these questions have not been answered in one coherent study analyzing the content of European news media. In that sense, this study is unique, because it is based on an extraordinary large data set: the European Election Campaign Content Analysis of 2009 i . For this project all news items in the TV evening news of the major newscasts on national television and least two newspapers per country ii in the three weeks before the European Parliamentary election of 2009 have been coded leading up to a database of more than 30.000 news items iii . This means, that the conclusions drawn are of higher generalizability compared to previous studies of European newsflows that are mostly based on data collected in not more than 5 countries. Moreover, having such an extensive data set at disposal is in fact a precondition for the analytic techniques utilized in this study. The lack of suitable data might be the reason that the European public sphere has never been conceptualized and empirically investigated as a network of interconnected national public spheres before, although this approach fits well to the phenomenon, at least when it comes to measuring the horizontal dimension of Europeanization of national public spheres. Knowing potential breaches in the European public sphere does not only satisfy scientific curiosity. There are also two important societal implications of systematic loopholes in the European network of newsflows. Firstly, the visibility of national actors and events in the European news media influences on the power structure of the EU. Which countries’ governments are being heard all across Europe and who does not get their say? Being visible to a greater public is a precondition for explaining ones standpoint and to solicit for support. The same holds true for national events and debates. Only those being picked up by news media of a majority of other member states can also exert an influence on the European agenda. Secondly, the fact that there are gaps in European newsflows itself is noteworthy. The European Union is a complex multi level system, in which developments in one member state can influence the whole entity. The initial Polish refusal to sign the treaty of Lisbon, and the subsequent negotiation process in which Poland was able to alter the treaty in its favor is a good example for the power of national players in the EU. However, in a patchy network of European newsflows most of the relevant national developments that motivate the behavior of national players in Brussels go unnoticed. This means, that a majority of the European 3

European Newsflows citizens is unable to give meaning to the positions and decisions of actors that have a national as well as a European function (e.g, the prime ministers). Thus, the present study investigates the structure of the European network of newsflows and its determinants. In the following we first sketch the academic debate concerning the European public sphere to classify the present study as an analysis of the horizontal Europeanization of national public spheres. Subsequently, we draw on news value theory and the empirical body of literature that tested and modified the initial ideas of Galtung and Ruge (1965) to propose an explanatory model. In a next step, we present a visual depiction of the network of European newsflows that is based on multi dimensional scaling of the mutual coverage in 25 European member states during the European Election campaign of 2009. Moreover an explanatory model of the structural features of the network, in particular the relative proximity of the member states and their integration in the network, is being suggested. Finally, we discuss the implications of these findings as well as future paths for European public sphere research.

4

European Newsflows

Theory  The emergence of a European public sphere  The academic debate concerning the existence and development of a common public sphere in the European Union has engaged scholars from all disciplines of the social sciences and is far from being closed (for an overview see Machill, 2006). Depending on the concept of European public sphere applied and the methods used to measure it opinions on whether and to what extent the European public sphere exists, differ significantly. The body of literature dealing with the subject is still growing: in the past ten years at least 1430 scientific studies that refer to the term ‘European Public Sphere’ have been published iv . Given the political relevance of the subject the considerable scientific effort devoted to the conceptualization and analysis of the public sphere of the European Union should be no surprise. The issue is of such significance in no small part because it feeds into the ongoing debate regarding the alleged democratic deficit of the European Union and its consequences for further EU integration (see e.g. Meyer, 1999; Scharpf 1999). Since the European Union has developed beyond mere intergovernmental cooperation of the member states into a set of supranational institutions entitled to take political decisions that affect the citizens of the member states, several scholars express concerns regarding the democratic legitimacy of policy making by European institutions. One of the key arguments made by advocates of the democratic deficit notion is that the European Union lacks a common public sphere, which is arguably a necessary condition for any functioning democracy. Firstly because, as Scharpf (1999) points out, political institutions like the EU lack legitimacy if they merely represent a group of individuals that do not share an arena for exchange of ideas and to inform themselves about political propositions. Moreover, a shared public sphere is also a precondition of any collective identity such as, say, a common European citizenship (Kraus, 2004). Kraus argues that if an aggregate of individuals that lacks a common identity is taking decisions by popular vote, a society in which minority interests routinely subjugated to the majority comes about. Finally, Koopmans and Erbe (2004) mention three more democratic functions of the European mass media, as the constitutive element of a European public sphere v , that are arguably not completely fulfilled: (a) The mass media should serve as a ‘response channel’ for political authorities, to inform themselves about the needs and interests of the people, (b) by covering relevant political events mass media make the polity structures and decision making processes transparent, (c) the mass 5

European Newsflows media provide the opportunity for all members of the political community to engage in the public debate (for example through interest groups and NGOs). The assessment of the extent to which those democratic needs are being met by the European public sphere in its current state might differ among scholars who study the phenomenon. There is, however, a general consensus that a transnational public sphere consisting of truly European mass media has not yet emerged. Van de Steeg (2006) even doubts whether it will ever exist, and calls a supranational European public sphere a utopia. Alternatively, the European public sphere is commonly conceptualized as a group of Europeanized, interconnected, national public spheres. That said, there is little agreement regarding the definition and correct operationalization of such a ‘Europeanized national public sphere’, and – depending on the conceptualization – there are remarkable differences in the evaluation of its existence. Gerhards (2000; 2002) defines ‘Europeanization of national public spheres’ as the attention national mass media devote to European subjects or actors compared to the coverage of national topics and politicians. In multiple studies concerning the development of the prominence of EU topics and actors over time in the German mass media he finds no evidence for an increase in European coverage in national news outlets at all. Accordingly, he draws the discouraging conclusion that “Europeanization of the public sphere lags far behind the Europeanization of politics” [translated by the author] (Gerhards, 2000 p.7). Trenz (2004), on the other hand, uses a comparable operationalization of Europeanization, namely the share of EU related coverage compared to other political coverage, but comes to a different conclusion. He finds sufficient evidence of political coverage using the EU as a referential frame, and highlights the importance of EU institutions as agenda setters within this coverage. Koopmans and Pfetsch (2007) and Koopmans and Erbe (2004) focus on the prominence of European actors and institutions in an analysis of claims in the German news media. They conclude that the importance of European actors in coverage depends on the topic. The more a policy field is Europeanized the higher the likelihood of European actors appearing in the news. Also opposing Gerhards’ results, they conclude that the Europeanization of the German public sphere accurately reflects the Europeanization of policy making. Conversely, Meyer (2005) argues that it is not only the amount of coverage that matters in determining the state of the European public sphere, but the assessment should also take account of the consonance in the fashion of sense making and framing in the European member states. And indeed several studies – qualitative and quantitative in nature – point to 6

European Newsflows the conclusion that the arguments and frames applied in national news outlets across Europe are quite comparable (de Vreese, Peter, & Semetko, 2001; Eder & Kantner, 2000; Meyer, 2005; Peter, 2003; Peter & de Vreese, 2004; Peter, Semetko, & de Vreese, 2003; Semetko & Valkenburg, 2000; Trenz, 2000, 2004). Nonetheless, two problems should be noted with regard to the validity of these studies. First, a number of them are case studies (e.g., Meyer, 2005; Trenz, 2000) of specific debates in only a few member states and therefore of limited use for drawing generalized conclusions. Second, the present studies of news coverage have not – with the exception of the work of van de Steeg –included non-European news outlets and cannot therefore exclude the not unlikely possibility that the consonance in frames and arguments are in fact the consequence of a global trend in the reporting of a certain issue. Next to analyses of the volume of coverage regarding EU institutions or topics in national news outlets (also called vertical Europeanization of national public spheres) and the argumentative consonance of national news media, a third dimension of Europeanization of national public spheres can be observed, as authors like van de Steeg (2006) point out, namely the degree to which the national media outlets notice relevant events and debates taking place in other member states (horizontal Europeanization). As national government officials determine most of the decisions taken by the European Union in the European council by applying the consensus principle, it is arguably relevant to be informed about political and social developments in all EU member states that determine the negotiation positions of the representatives of national governments involved in the Council. However, until now there have been only a few empirical cross sectional studies of the communicative linkages between member states, which are mostly of descriptive nature. Kevin (2003), for example, compares the relative and absolute coverage of the affairs of other member states in 15 countries, identifying Sweden, Austria, Germany and the Netherlands as the most ‘Europeanized’ public spheres. There has been, however, extensive research with regard to the horizontal Europeanization of the German public sphere. Koopmans and Pfetsch (2007) analyse the claim structure of the German media, and conclude – similar to the results of Meyer (2005) – that the degree of horizontal Europeanization depends on the Europeanization of the policy area in question. Agricultural and monetary political issues are discussed with reference to other member states more often than problems in policy fields that are still primarily regulated by national authorities. With regard to the significance of actors representing other member states, Trenz (2004) determines, in a broader study of the Europeanization of the German public sphere, 7

European Newsflows that 12% of coverage dealing with the EU is driven by EU government officials not situated in Germany. Nevertheless, a large scale, cross sectional, elaborate study of the newsflows between the European member states has not yet been conducted. This is probably due to a few empirical and methodological obstacles. The first of these are the problems arising from the precondition to conducting valid and reliable content analyses in at best 27 member states: cultural differences between coders, translation issues and incomparability of news outlets, to name just a few. Secondly, there is no benchmark with regard to the amount of coverage that would be sufficient to call a national public sphere Europeanized. How much do we actually need to know about the events and debates that take place elsewhere in Europe? Finally, there is also a statistical problem: Since there is only limited space in the foreign news section of newspapers and TV newscasts the prominence of one country influences the likelihood of other countries making it into the news. If there is, for example, a major event in the UK, there is little room to discuss current developments in Southern Greece. This means that the observations are not independent of each other, a precondition for most types of explanatory analyses techniques. Despite the empirical and analytical hurdles, this study aims to map and explain the newsflows in the European Union to visualize and understand the processes of horizontal Europeanization. The data used has been obtained by the European Election Campaign Study through content analyses of national news media in 25 of the 27 member states. The problem of an undefined benchmark for horizontal Europeanization is addressed by using the Europeanization of other member states as point of reference. Moreover, the focus of this study is not merely to assess the absolute or relative volume of coverage of other member states in one EU country but rather to identify clusters and gaps in the European news network by applying social network analysis techniques. Network analysis not only permits a comparison of the integration or lack thereof of member states in the network of newsflows, but also the visualization of the structure of the network to ultimately arrive at a comprehensive map of the horizontal Europeanization of national public spheres.

Determinants of international newsflows  In order to explain and understand the attention which the mass media in one member state devotes to events in other member states I draw on news value theory. According to this approach, first described by Walter Lippman (see Kepplinger, 1998), the mass media function as gate keepers who systematically select from an infinite number of events that could 8

European Newsflows potentially be covered, those that show the highest cumulative news value (in terms of relevance and appeal for the audience) (see also White, 1950). Galtung and Ruge (1965) translate this general approach into the context of foreign news coverage to highlight and explain the asymmetric attention news outlets all over the world devote to events in foreign countries (see also Harcup & O'Neill, 2001). According to the authors, countries that belong to the Western hemisphere are generally overrepresented on foreign news pages. Their argument is linked to the cultural hegemony debate, initialized by the work of Schiller, that took place in the 1970s and critically discussed the global cultural dominance of the US. In their case study of the coverage of three foreign conflicts in the Norwegian press, Galtung and Ruge provide empirical evidence suggesting that geographic, economical, and social characteristics of a country indeed determine the amount of coverage it attracts in foreign news outlets. Moreover, they conclude that the characteristics of the country covered also influence the kind and slant of news the audience abroad receives. The less culturally, economically and geographically close a country is to another, the more violent and negative are the news appearing in national news outlets. However, as their study is not representative in the sample of countries and conflicts, and the operationalization of key independent variables, such as the elite status or cultural proximity of a country, remains vague, the empirical evidence for their theory is less convincing than their theoretical ideas. The notion that the characteristics of the country covered determine the extent to which it appears in the news has been picked up by many scholars in the social sciences and lead to the fruitful debate concerning the geography of news that took place in the 1970s and 1980s. In order to investigate the degree mass media are truly dominated by a few elite nations Rosengren & Rikardsson (1970) use extra media data to operationalize the independent variables. This means they use external sources, in particular statistics of international trade, and population size in explaining the variance in foreign coverage. The authors conclude that, powerful countries – in terms of foreign trade and population size – are indeed noticed more frequently than others. This has been an important step in news flow research, as Galtung and Ruge relied on information gathered through content analysis only (Rosengren, 1970; Rosengren & Rikardsson, 1974). Besides Rosengren many other scholars conducted research using a variation of independent variables (for an overview see Wu, 2000). In an effort to systematize the extended body of literature, Wilke (1998) proposes a typology of 7 dimensions of relevant macro level variables that explain the variance in international newsflows. These dimensions can be further distinguished into variables which describe the characteristics of the country covered, and dyadic variables defining the proximity between 9

European Newsflows the covered and the covering country, though not all of them have received an even amount of attention in academic research (see Table 1).

10

European Newsflows Table 1: Dimensions of determinants of international newsflows  Dimension

Independent Variable

Empirical studies (selection)

Geographic features

Size

(Wilke, 1998)

Geographic proximity (dyadic)

(Nnaemeka & Richstad, 1980; Westerstahl & Johansson, 1994)

Political and economic system

Location (continent)

(Wanta, Golan, & Lee, 2004)

GDP / GDP per capita

(Hagen, Berens, Zeh, & Leidner, 1998; Ishii, 1996; Rosengren & Rikardsson, 1974)

Trade (dyadic)

(Ahern, 1984; Robinson & Sparkes, 1976)

Colonial ties (dyadic)

(Atwood, 1985; Galtung & Ruge, 1965; Skurnik, 1981)

Infrastructure

Presence of satellites, news

(Ishii, 1996; Wu, 2003)

agencies, foreign correspondents Media system

Methods of news gathering,

(Wallis & Baran, 1990)

number of news outlets (characteristics of the covering country) Journalism

Role perception of journalists

(Wallis & Baran, 1990)

(characteristics of the covering country) Characteristics of the audience/

Common official language

(Hagen et al., 1998; Kim & Barnett,

cultural proximity

(dyadic)

1996)

Diplomatic proximity (dyadic)

(Westerstahl & Johansson, 1994)

Valence, Surprising, Consonance,

(Galtung & Ruge, 1965)

Characteristics of the event

Frequency

The drawback for most of these studies, however, is the limited sample size on the macro level. With only a handful of countries under investigation, the variance in the independent variables is so restricted that a fully reliable explanatory analysis is not possible. Larger studies of ten or more countries, on the other hand, struggle with the question of how to 11

European Newsflows validly measure the dependent and independent variables in a multi-cultural context. To give an example, Kim and Barnett (1996) use international circulation of national newspapers as an indicator for global newsflows. The validity of this measure can be doubted though, as it might not necessarily indicate the newsflows but rather the size of the expatriate migrant group interested in news from their country of origin. One of the few large scale, explanatory studies of international news flows is the study by Hagen et al. (1998). The authors use data collected in the second UNESCO coordinated foreign news study to identify the determinants of international newsflows, operationalized as the share of news items about country A in country B compared to all foreign coverage. The authors find two significant factors predicting the news flow between the two countries: The first is a characteristic of the country covered, namely its power status. The authors operationalize power status as a combination of economic power (indicated by its GDP), military power (budget for defense), and scientific power (number of scientific publications per year). In their study they show that the greater the power status of a country the higher is its chance of being recognized by international news media. This correlation particularly holds in the EU. Secondly, they identify a number of dyadic variables of significance to the prediction of international newsflows, implying that not only do the characteristics of the country that is being covered matter in determining the amount of coverage, but also a combination of characteristics of both the covering and the covered country, among those quite prominently the share of trade volume, a common official language and socioeconomic similarity. This notion is in line with theoretical advances in the 1980s and 1990s. At this point world system theory, which proclaims the domination of Western countries in the global cultural flows was criticized for not acknowledging the cultural exchange taking place in geographic centers like South America, Europe, or South Asia (Hardy, 2008). Accordingly, the global structure of newsflows was no longer understood as a one-way street in which the powerful dominate the smaller and poorer countries, but rather as a network of unilateral and multilateral flows in which globalizing as well as localizing mechanisms are of significant importance (Nnaemeka & Richstad, 1980). As a result, regional clusters of countries that intensively report about each other have been identified, for example in Africa (Atwood, 1985) or in the South Pacific region (Nnaemeka & Richstad, 1980). But as noted above, the horizontal communicative linkages within Europe have not been in the focus of academic research yet.

12

European Newsflows

Hypotheses  Not all of the proposed independent variables are of equal importance when it comes to explaining the patterns of horizontal Europeanization of the national public spheres of the member states. Colonial background, for example, is of no particular relevance and differences in infrastructure (presence of international news agencies, satellites, transmission opportunities) are arguably too small to have a significant impact on the European newsflows. vi So, what are the factors that shape the structure of inter-European newsflows? To answer this question, it is helpful to distinguish two network characteristics a) the volume of newsflows between two member states and b) the centrality of a country in the network. Whereas the first focuses on the relation of two nodes in the network, the second characteristic accounts for the position of a country with regard to the structure of the entire network. Volume of news flow between two countries  According to Galtung and Ruge (1965), the relative share of newsflows between two countries depends on their power status and proximity. In previous research four dimensions of the phenomenon proximity have been suggested: cultural proximity, geographic proximity, historical links, and economic proximity (see Table 1). Validly measuring cultural proximity is a challenging task. The indicators applied in previous research reach from staff members employed in an embassy abroad (Westerstahl & Johansson, 1994) to socioeconomic similarity (Hagen et al., 1998). In this study, affiliation to a common language family is chosen as an indicator for cultural proximity. In Europe belonging to the same language group implies that the people of the concerned countries formed a cultural entity at an earlier stage in history. Some of those linguistic clusters still demonstrate a collective identity, e.g. Scandinavia; Denmark, Sweden, and Norway. These three countries, which form the Northern European language family cooperate closely politically as well as economically vii . Moreover, the languages within one cluster are in some cases so similar as to be mutually intelligible (e.g. Slavic languages, Romantic languages), which has eased cultural exchange through the centuries. Accordingly, it can be assumed that countries belonging to the same linguistic group report more about each other. H1: The volume of newsflows within the same linguistic cluster is larger than between different linguistic clusters.

13

European Newsflows In previous research geographic proximity has been shown to have no significant influence on global newsflows (Hagen et al., 1998). Due to its proclaimed theoretical relevance (Galtung & Ruge, 1965) however, it is included in this study nevertheless as another potential explanatory factor for the intensity of newsflows between countries. H2: The smaller the geographical distance between two countries, the greater the volume of newsflows. There are many alternative ways to systematize the historic relations in Europe. The reasoning behind the inclusion of measures of historic proximity in an explanatory model of international newsflows is that past political alliances have influenced the economic and cultural relations between states in question. In previous studies of international newsflows historic proximity has often been operationalized as colonial background (e.g., Atwood 1985). However, as pointed out earlier, colonialism as a measure of historic proximity is of little relevance in the European context. In this paper, we use a different approach, namely to study the impact of the most recent division of Europe by the Iron Curtain as a measure of historic proximity. This division of Europe could be a significant factor in explaining the structure of the network of newsflows in Europe, since cooperation between the blocs was kept to a minimum during the period of the Cold War, whereas countries on either side of the Iron Curtain were encouraged to interact with one another. This arguably had an effect on the economic and cultural proximity between countries of the two blocs and therefore on the perceived relevance of events taking place in the respective countries. H3: Countries of the former Communist bloc / West European bloc are more likely to cover countries from the same bloc. Finally, the economic relations between two countries themselves should be considered in order to explain their mutual newsflows. The more intensively a state economy is linked to a foreign state economy, the more relevant events in this particular country become (Hagen et al., 1998). H4: The greater the exchange of goods and services between two countries, the greater is the volume of newsflows. Centrality  The second phenomenon is, in network analysis terminology, a “node characteristic”, namely the centrality of a node in the network. Centrality depends – according to news value theory – on the economical and political power status of the country. The more important a country 14

European Newsflows appears to be, the more attention will the mass media in other countries devote to events taking place in that particular country. This influences its integration in the network in a positive manner. In accordance with previous research (e.g. Hagen, et al., 1998; Ishii, 1996; Rosengren & Rikardsson, 1974) the following hypotheses can be formulated. H5: The greater the economical power of a country, indicated by its GDP, the greater is its centrality in the European news network. H6: The greater the political power of a country, indicated by its contribution to the EU budget and the number of representatives in the EU parliament, the greater is its centrality to the European news network. Besides these traditional determinants of foreign coverage as introduced and discussed above, we are interested in the predictive power of actual integration in the EU with regard to the integration of a member state in the European news network. To our knowledge, the influence of regional political cooperation on newsflows has not been tested before. However, it can be argued that the efforts of the European Union to integrate member states economically and culturally affect the cultural and economical proximity of the members over time. Hence, the longer a country participates in the European Union, the better it is connected to other member states and therefore increasingly central to the European network of newsflows. H7: The longer the membership of a country in the European Union, the higher is its centrality in the European news network.

Method  In order to investigate the structure of the horizontal communicative linkages between European member states, a network of newsflows is mapped using Multidimensional Scaling (MDS), analyzed by applying the tools of social network analysis and explained using regression analysis.

Sample  The analysis is based on content analysis data provided by the European Parliament Election Campaign Study as part of the PIREDEU framework. In this cross sectional research all news items on the front page, a randomly selected additional page, and all stories about the EU viii from at least two major national newspapers, as well as all news items in two national TV news broadcasts from all European member states have been systematically analyzed ix . 15

European Newsflows However, the Danish and Finnish data is not considered in this paper, as the data collection has not been completed in these countries yet. The field period included the 21 days before the European parliamentary elections in that particular member state, which means that there is a two to three day deviance between the countries, as the elections were spread over a period of three days across Europe. Observing horizontal Europeanization during a parliamentary election campaign instead of a random point in time has two distinctive advantages: 1) The general amount of EU coverage is generally higher during periods preceding elections x (see Gerhards, 2002), which means that the phenomenon can be observed more easily and 2) the Parliament is constituted by popular vote in all member states, implying that the political developments in all member states that influence voter choice are of comparable relevance, which reduces the chances of an event driven bias. The disadvantage of this choice of field period, on the other hand, is that the results cannot be generalized to other periods. The content analysis was conducted by 77 coders trained and supervised by academic staff in Amsterdam (NL) and Exeter (UK). To assure intercoder reliability all trainers participated in an inter-coder trainer reliability test that yielded satisfactory results and two of the Amsterdam coder trainers were present during the coder training in Exeter. The intercoder reliability with regard to the variables used in this study is reasonably high. In a reliability test conducted before the start of the coding Krippendorfs alpha was 0.70 for the Exter group of coders and 0.65 for the coders located in Amsterdam for the main dependent variable in this study.

Operationalisation  Dependent Variables  The network of newsflows is constructed by applying Multidimensional scaling calculated on basis of the amount of stories published in country A that cover events country B xi as a measure of proximity. To account for the variation in outlet length and number of outlets per country the data is weighted to create an even amount of news stories per country for the purpose of this study.

16

European Newsflows

Table 2: Number of news stories by covering country (weighted) Austria Belgium Bulgaria Cyprus Czech Republic Estonia France Germany Greece Hungary Ireland Italy Latvia Lithuania Luxembourg Malta Netherlands Poland Portugal Romania Slovakia Slovenia Spain Sweden UK

AT

BE

BU

CY

CZ

EE

F

GE

GR

HU

IRE

IT

LV

LT

LUX

MT

NL

PL

PT

RO

SK

SI

ES

SE

UK

X 6 10 0 15 0 50 176 4 21 13 44 2 4 2 0 13 19 0 8 8 0 25 2 67

4 X 1 0 3 0 158 47 2 1 7 39 1 7 4 1 49 5 1 0 2 1 11 2 66

7 24 X 1 13 0 40 30 12 0 6 28 4 12 1 1 6 1 0 9 0 0 12 0 56

0 28 0 X 17 0 52 20 188 2 11 56 4 6 0 0 7 4 2 0 22 2 7 6 169

3 4 0 0 X 0 18 12 1 4 1 6 0 0 1 0 5 5 0 1 12 0 2 0 20

15 12 0 0 15 X 44 33 3 0 12 18 36 36 0 0 6 9 0 0 0 0 9 9 65

6 17 2 2 11 5 X 34 3 5 17 50 6 3 3 2 20 14 3 2 2 0 28 14 103

18 8 3 0 14 1 60 X 3 10 10 37 5 12 14 1 15 22 0 3 1 0 32 5 103

1 17 1 13 1 0 44 46 X 0 6 29 1 0 3 0 8 0 1 0 0 0 9 3 66

19 14 4 2 12 2 58 49 6 X 2 35 0 2 2 0 12 17 2 45 39 0 25 6 68

2 12 6 0 4 0 38 26 0 2 X 34 6 2 0 0 6 6 2 0 0 0 12 0 218

2 5 0 0 2 0 54 117 0 0 3 X 2 0 2 0 8 0 0 0 2 0 11 0 55

5 7 2 2 9 36 66 34 2 2 11 14 X 50 0 0 7 9 0 0 0 0 14 25 66

3 6 6 0 22 9 40 37 0 0 9 16 37 X 0 0 6 28 6 6 0 0 0 3 50

36 59 12 14 10 5 133 147 5 5 12 40 2 14 X 2 24 17 5 7 2 5 26 5 52

2 8 2 9 3 3 14 17 3 2 6 33 5 6 0 X 2 5 0 0 2 0 3 3 62

5 34 0 0 5 0 67 95 4 0 13 34 2 7 0 0 X 25 0 0 0 0 16 2 103

0 10 6 0 17 2 42 67 4 4 12 52 6 12 0 0 8 X 0 10 8 0 17 8 71

7 15 2 0 5 0 38 13 0 0 8 28 0 0 0 0 10 10 X 2 0 0 32 0 48

6 26 5 0 8 0 66 16 5 10 2 23 0 6 2 0 2 6 3 X 2 0 15 0 39

17 79 10 0 138 2 74 21 5 12 10 26 2 10 0 0 10 24 0 2 X 0 7 5 57

36 3 5 2 10 5 22 75 3 8 3 58 0 2 2 0 8 12 0 0 2 X 15 5 78

16 10 3 1 6 1 70 34 2 2 10 47 0 0 1 0 9 5 5 1 1 1 X 3 78

2 9 9 0 0 4 49 33 4 7 0 22 16 7 0 0 13 9 0 0 4 0 2 X 63

2 0 0 2 5 0 59 31 5 0 16 28 2 2 0 0 9 7 7 0 0 0 14 0 X

17

European Newsflows The centrality of the member states in the network is assessed by calculating the flow betweenness of each country. Flow betweenness is a node attribute that measures the number of times a country is on the shortest path between two other countries weighted by the tie strength. This measure accounts for the position of a node in the network relative to all other nodes and hence the structure of the whole network. Flow betweeness is a popular indicator for integration in a network in social network literature (Hanneman & Riddle, 2005). Independent Variables  Data concerning the country attributes serving as independent variables in the analysis: volume of trade xii , geographic proximity xiii , affiliation to Eastern or Western Europe xiv , language family parliament

xviii

xv

, GDP

xvi

, population

xvii

, number of members in the European

, length of EU membership, and contribution to the EU budget xix has been

collected from several official and academic sources. Following the Rosengren argument that an explanatory variable of international newsflows should not rely on content analysis data but rather on objectively observed country characteristics, all independent variables are operationalized on the basis of extra media data.

Analysis  Descriptive Analysis  The network is constructed on the basis of Multidimensional Scaling. This analysis technique transforms an adjacency matrix of a group of elements into a visual configuration of those elements according to their relative proximity. The configuration of the countries in the network of European newsflows is based on a matrix of mutual newsflows between two countries. This means that the closer two countries are located in the network, the greater is their relative exchange of newsflows. Each node in the network represents one of the 25 member states under investigation. The nodes are connected by dyads that represent the newsflows. The thicker the line that connects two countries, the larger the amount of stories published in each country about events that take place in the other. Explanatory Analysis  As noted above, one of the fundamental assumptions of OLS regression analysis - the independence of observations - is violated in the case of the data under investigation in this

18

European Newsflows study. Consequently, the explanatory analysis is calculated using QUAP (quadratic assignment procedure) modeling specifically designed for network analysis. The method accounts for the restrictions of network data (see Nagpaul, 2003). To address the autocorrelation in the data, the significance levels of the effects are calculated by bootstrapping. This means that the rows and columns of the dependent matrix are permuted randomly in multiple analyses to create a random distribution which is used to test the significance of the model and the estimators.

Results  Descriptive analysis  Figure 1 shows the network of European newsflows during the European Parliament election campaign xx . The density of the network depicted is rather high: 93% of all potential links between the member states exist in the symmetrized network. However, if only those links that make up at least 3% of the share of all stories about other member states of the covering countries are considered, the density is much lower. In that case only 33.8% of all potential links are actually present in the network. This alludes to an asymmetric amount of attention payed by national news outlets to other European member states. And indeed, when looking at the configuration of the nodes in the networks, it becomes obvious that the four largest member states (Germany, France, the UK, Italy) are located at the heart of the network and appear to be well connected to all other member states. Around the center of the network we find four states in intermediate distance to the midpoint: Ireland, Belgium, Luxembourg and the Netherlands. The latter three, and three of the center countries (France, Germany, Italy), are also the founding members of the European Union. All other member states are located in the periphery of the network, mostly structured according to their cultural and geographic proximity to each other. Remarkable is the separation of the periphery in Eastern and Western Europe. If a horizontal line is drawn through the middle of the network, almost all Eastern European countries are located on the right of the network and the Western European countries can be found on the left. There are, however, a few noteworthy exceptions. Slovenia is located between Spain, Italy, and Austria; and is connected through stronger newsflows to Western European countries only. The news outlets of Slovenia, as the only Eastern European country participating in the common currency Euro, seem to have a clear focus on Western Europe. Austria, on the other hand, is placed in the middle with regard to the East/West dimension. 19

European Newsflows Figure 1: Network of newsflows (all stories) Bulgaria

Slovenia Austria Spain

Hungary Luxembourg

Italy

Greece

Germany

Poland

Cyprus

France UK

Slovakia Belgium Czech Republic

Netherlands Malta

Ireland Lithuania

Latvia

Portugal Estonia Sweden

Romania

With regard to the relative proximity of countries in the periphery to each other, it can be seen, that there are a few excpetionally tight links: these are between Slovakia and the Czech Republic, Lithuania and Latvia, and Greece and Cyprus. The newsflows depicted in Figure 1 are calculated on the basis of all news stories in the sample. Figure 2, on the other hand, depicts the network of newsflows, if only those news stories that explicitly refer to the European Union are considered.

20

European Newsflows Figure 2: Network of newsflows (EU stories) Malta Romania Slovenia

Slovakia

Czech Republic

Sweden Poland

Hungary

Belgium Spain

Austria Germany

UK

Cyprus

Lithuania

Greece

Luxembourg France Latvia

Netherlands

Italy Bulgaria

Ireland

Portugal

Estonia

Compared to the network calculated on the basis of all stories there are only a few small differences. Poland has moved up to the group of countries that surround the center, implying that the country is more central to the network if only EU stories are considered. Also, the Eastern focus of Austria is even more obvious in the EU stories network. The country is not longer located on the border between East and West but can be now found within a group of East European countries. Interestingly, Portugal is also located among the Eastern European countries. Except for the link between Cyprus and Greece, the smaller regional clusters seem to be of less importance when it comes to the structure of newsflows. The distance between Latvia and Lithuania, as well as that between Slovakia and Czech Republic relative to other links in the periphery of the network, is bigger in the second, EU-specific network.

21

European Newsflows

Explanatory analysis  Newsflows  Due to the potential multicollinearity of the independent variables, in particular language family and affiliation with Eastern Europe (a majority Post-Communist countries belong to the Slavic language group), several models have been calculated to estimate the effect size and significance of the independent variables (see Table 3) predicting the volume of newsflows between member states. The first model includes trade relations, geographic distance, and the East/West affiliation; in the second model the East/West affiliation is replaced by language family affiliation. All of the above mentioned variables are included in the third model, and the last model yields the results of a regression in which two key attributes of the country covered (economic strength and population size) are controlled for. As it becomes obvious from Table 3, East/West affiliation as well as language family affiliation are significant if tested independently of each other. Belonging to the same language family increases the predicted number of news stories by 8.1, sharing the same political past accounts for 2.6 more news stories. This means that Hypotheses 1 and 2 are supported by the analysis. However, if both factors are tested in the same model, the East/West affiliation becomes insignificant. As pointed out earlier, this arguably due to the high multicolliniarty of language group and East/West affiliation. Therefore, the first model, and not the third model should be considered here. Both trade and geographic proximity are significant in all models presented, implying that Hyptheses 3 and 4 can also be accepted. A closer look at Model 4 reveals, however, that the GDP of the country covered accounts to a large extent for the variance explained by mutual trade. That being said, the influence of trade is still significant, even if the economic strength of the country that is covered is controlled for. Table 3: Unstandardized QUAP regression coefficients predicting dyadic newsflows Trade (in billions) Geographic distance Eastern Europe Language family GDP1 (covered c.) Population (covered c.)2 Intercept Adj. R2

Model 1 0.86** -0.003* 2.60*

14.89 0.171**

Model 2 0.82** -0.003* 8.10**

Model 3 0.81** -0.003* 1.28 7.64**

Model 4 0.15* -0.004** 2.14 8.65** 3.65** -0.35

14.93 0.179**

14.16 0.178**

7.10 0.488**

2000 permutations, N=600; *:p<0.05; **:p<0.01; 1 in 100 quadrillions, 2 in millions

22

European Newsflows Centrality of member states  With regard to the explanation of the integration of the individual member states into the network the results are less obvious (see Table 4). Taken by themselves (Model 1-4), all independent variables are of statistical significance when it comes to explaining the centrality of a node in the network. This supports the results of the visual examination of the news network. Table 4: Unstandardized regression coefficients predicting flowbetweeness of EU memberstates Population (in mio) GDP1 EU MPs EU integration Budget contribution

Model 1 0.38*

Model 2

Model 3

Model 4

Model 5

0.002**

Model 6 -1.27 4.01 0.52 0.20 -0.0004

1.53* 0.36* 0.45**

Intercept F test

19.40 5.844*

23.15 7.802*

16.78 5.821*

16.61 7.056**

20.11 7.205**

11.62 1.862

Adj. R2

0.136

0.191

0.135

0.171

0.175

0.128

1

1000 permutations, N=25; *:p<0.05; **:p<0.01, in 100 quadrillions

However, if all factors are tested simultaneously, none of them is significant. Moreover, two of the factors (population and budget contribution) are predicted to have a reversed effect direction. Nevertheless, given the small case number (N=25), and the multicolleniarity of the independent variables the reliability of the inclusive model is arguably low. Therefore, we can cautiously conclude that hypotheses 5 to 7 are supported by the analysis, acknowledging that some of the correlations found might be spurious.

Conclusion and discussion  In the discussion of the theoretical aspects of this study it was suggested that the horizontal Europeanization of national public spheres of EU member states is an essential dimension of the European public sphere today. As such, it is essential to understand the determinants that shape the patterns of newsflows across Europe. Which countries are given attention by the mass media of other European member states? And, potentially even more relevant, where are the gaps in the network of European newsflows? News value theory, in particular the strain of academic literature that focuses on selection and publication mechanisms in foreign news, has been proposed as a theoretical framework to understand the structure of the network of European newsflows. An inventory of theoretical and empirical work in this field has lead to the identification of two clusters of factors that 23

European Newsflows determine newsflows between countries: a) determinants measuring the economical and political power of the countries covered and b) measures of the proximity (geographic, cultural, economic, and historic) of the covering and covered country. The findings of the present study demonstrate that all of the suggested factors are of relevance in explaining the structure of the network of newsflows in the European Union. The results imply that, although the network of newsflows appears to be very dense at first glance, the centrality of a country in the network and the volume of newsflows between two member states depends to a large degree on the factors mentioned above. Moreover, if the weak links, representing newsflows of less than 3% of the coverage of other EU member states, are disregarded, the network of newsflows is in fact quite patchy: only one third of the potential links in the network actually exist in this case. In other words, events in a majority of the European member states go unnoticed in a large part of Europe. But what does it mean if events or debates taking place in one of the smaller, less powerful countries are being ignored by the mass media in almost all other EU member states? To give a specific example, over a period of 3 weeks, there were only 6 news stories about the Baltic countries in the Austrian news outlets in the sample. At the same time there were 176 news items about events transpiring in Germany. This example clearly demonstrates that the asymmetric representation of the world in foreign news coverage, first criticized by Galtung and Ruge (1965) and repeatedly observed in subsequent empirical research (e.g., Ahern, 1984; Hagen et al., 1998; Ishii, 1996) also applies in the European case. Even within the Western world the strong economical and political players are covered much more prominently in the foreign press. And indeed, as the analysis in this paper has demonstrated, the UK, Germany, France, and Italy can be found at the heart of the European network of newsflows. What does this mean for the state of the European public sphere and its implications for the functioning of democracy in the European Union? The answer to this question is twofold. First, if we are only exposed to the ongoing debates and events of the largest member states and are not informed about relevant developments in the other member states, we lack understanding of the motivations and positions of the representatives of those member states at the European level. And even though we are not eligible to vote for the members of the European parliament for that particular country, nor the political staff that is sent to negotiate in the European Council, their decisions impact policy making for the whole of Europe. Being 24

European Newsflows uninformed about the problems debated in other member states also precludes any opportunity to participate in the discussion that is leading up to a decision in the end. The second implication we would like to allude to is the bias of the news favoring the most powerful countries. If the arguments and pronounced interests of the political staff of the largest political countries are being heard all across Europe, the voices of the most powerful political players become doubly amplified. This raises concerns with regard to the representation of the interests of smaller countries. If they are not debated in a significant part of the European public sphere, the chances of political proposals supporting the interests of the smallest member states entering the European policy cycle are arguably lower. Thus, Kraus’ (2004) concerns with regard to an underrepresentation of minority interests due to the lack of a common European public sphere, are in fact legitimate, if only the horizontal Europeanization of national public spheres is considered. A second interesting finding is the significance of cultural and geographic proximity to the volume of newsflows between two countries. In previous research (for example Hagen et al., 1998) geographic proximity has been dismissed as insignificant to the variance in newsflows between countries. In the light of the present results, this might be due to some strong intercontinental links (for example USA – Europe). If the structure of foreign news within one continent is being investigated, as in the case of the present study, the geographic distance turns out to be of statistical significance. Theoretically, this is quite comprehensible. Events taking place in a country close by (for example an environmental hazard, or economical downturn) that might influence local developments are arguably more relevant to the audience in a neighboring country. There is, however, an alternative explanation for the correlation. Geographic distance could also be understood as a proxy for cultural or historical proximity, as most neighboring countries in Europe share a long history of military and cultural exchange. Therefore, this factor could illuminate a dimension of cultural proximity that is neglected by the other two indicators of cultural/historical proximity included in the regression analysis in this study: Language family and affiliation to post-communist or Eastern Europe. Language group affiliation indicates cultural bonds that date back far into history and therefore indicates the long term dimension of cultural proximity. Being a postcommunist country or not, on the other hand, is a rather short term attribute in terms of development of culture. If the cultural exchanges over the past centuries that go hand in hand with a shared border are interpreted as an indicator of intermediate-term cultural proximity in Europe, the factor geographic proximity fills a gap in a longitudinal operationalization of cultural proximity. 25

European Newsflows The division of the European news network into East and West, especially when it comes to the periphery, is another interesting finding of this study. Looking at the network of newsflows, we might even wonder whether we should actually speak of two European public spheres: an Eastern and a Western European. There are, however, two noticeable exceptions. Slovenia, a post-communist country, which has adapted rapidly to the economic standards of Western Europe and is as the only Eastern European country allowed to participate in the common currency, is strongly linked to the Western domain when it comes to newsflows. Austria, on the other hand, as a country that has deep cultural and economic ties with Eastern Europe (established during the Habsburg monarchy), has strong communicative links with the East. So, does Western Europe have to rely on Austria to explain the developments in the post-communist member states as virtually the only bridge to the East in the European network of newsflows? Arguably yes, if the present state of integration of the national public spheres is in question. But the division of Europe with regard to newsflows might not be permanent. The duration of membership in the EU has proved to be a significant factor in explaining the centrality of a country in the network. Since all the post-communist countries have joined the EU in the past decade, they might become better integrated in the years to come. In this study novel methods have been applied to generate a deeper understanding of the state of horizontal Europeanization of national public spheres of the member states. By applying network analysis techniques, it has been possible not to only create a visualization of the newsflows but also to provide a valid method to assess the integration or centrality of a country in the network. That being said, the method also has a few pitfalls. First of all, it is impossible to do complex modeling applying network analyses software. The tools to do network analysis are still being developed and are not integrated in standard statistical packages yet. This means that there are also limitations with regard to testing relevant assumptions for regression analysis such as the absence of multicollinearity, which raises concerns about the validity of the inferential statistics presented in this study. It can, however, be argued that testing the statistical significance of the results is not as important in this particular case, as no inferences for a general population are being made. Since almost the whole population of countries is included in the dataset, the study does not attempt to generalize on the macro level. And, as previously noted, as the period in time that is being investigated has not been chosen randomly, neither is the dataset representative for a specific time phase. That being said, the development of horizontal Europeanization over time might be an interesting aspect to investigate in the future. How does the network of 26

European Newsflows newsflows develop over time? Which historical/political events cause structural changes in the network? Another worthwhile focus in future research of the inter-European newsflows between member states could be to look at the prominence of actors from other EU member states in the national news. How often do they appear? Can they make their voice heard? And how are they evaluated? In this study the coverage of other European member states in the period leading up to the European elections has been investigated. The discouraging conclusion is that even in the context of such a major European event involving all member states evenly the network of news is rather patchy. When it comes to the degree to which the mass media notice events in other member states, Europe appears to be divided: between East and West, and between a few economically and politically powerful countries that are subject to coverage across Europe and a majority of member states which are only noticed by mass media in their neighboring countries. Hence, the horizontal Europeanization of national public spheres can be considered to still be at a rather low level, or to put it bluntly: we found hardly any evidence for the existence of common and coherent European public sphere at all. Arguably, the upward trend in the Europeanization of national public sphere with regard to vertical Europeanization, as established by work of Trenz (2000), Koopmans (2004), or Meyer (2005) does not hold for the horizontal dimension of the phenomenon. This implies that the European citizen might become increasingly familiar with the European polity setup and politics taking place in Brussels, but remains ignorant of the national events and debates that motivate the position of representatives of the governments of other member states at the European level. Besides its relevance for the European public sphere debate, this study also contributes to the discussion concerning the structure of global newsflows. Previous research in the field often takes on a global perspective, either arguing that a group of countries dominates the rest of the world, or, on the contrary, that the global newsflows are multilateral and news is organized in regional centers (Wu 2003). In the present study we have put the microscope on such a cluster, namely Europe, and established that the exchange of news within a cluster is far from being homogenous. Instead we found, that the very same mechanisms that determine global newsflows are also of relevance in explaining the asymmetries in regional newsflows. There is, however, an interesting difference: Cultural and geographic proximity is of much higher importance, compared to the explanatory power of economic factors. Geographic proximity was even found to be of no statistical significance in global newsflows, whereas we found a high level of significance of the factor. Potentially, this is a consequence of smaller differences in economic and political power within a region compared to the global context. 27

European Newsflows This means, that the particular nature of the network under study needs to be taken into account. When the power relations between countries are balanced, the mutual interest of two countries depends on their cultural and historic relations. Thirty years ago Rosengren and Windahl (1989) published a study that investigated the cultivation effects of an asymmetric representation of the world in the media. The authors asked Swedish school children to name important countries of the world, and found that most children knew the United States, and Europe, but hardly anybody mentioned African or Asian counties, especially those children who were frequent users of the media. What would the map of Europe look like, if we asked school children all across the continent to draw it today? Based on the findings of this study we can assume that the map would differ from state to state. Each time it would contain the most important countries: Germany, UK, France and Italy. But apart from that we would find that the ‘picture in the heads’ citizens have of Europe is as scattered as the depiction of Europe we find in our newspapers every day. The portrait of Europe is a mosaic, like the Union itself.

28

European Newsflows

References  Ahern, T. J. (1984). Determinants of Foreign Coverage in U.S. Newspapers. In R. L. Stevenson & D. L. Shaw (Eds.), Foreign News and the New World Information Order. Ames: Iowa State University Press.

Arends, E. (2009, 29.4). Babes voor Brussel: mevrouw Berlusconi boos. Volkskrant online. Atwood, L. E. (1985). Old Colonial Ties and News Coverage of Africa. Paper presented at the International Communication Association annual convention. De Vreese, C. H., Peter, J., & Semetko, H. A. (2001). Framing Politics at the Launch of the Euro: A Cross-National Comparative Study of Frames in the News. Political Communication, 18(2), 107 - 122. Eder, K., & Kantner, C. (2000). Transnationale Resonanzstrukturen in Europa. Eine Kritik der Rede vom Öffentlichkeitsdefizit. In M. Bach (Ed.), Die Europäisierung nationaler Gesellschaften (pp. 306-331). Wiesbaden: Westdeutscher Verlag. Flückiger, P. (2008, 1.7.). Warum Kaczynski plötzlich auf die EU-Bremse tritt. Welt Online. Galtung, J., & Ruge, M. H. (1965). The Structure of Foreign News: The Presentation of the Congo, Cuba and Cyprus Crises in Four Norwegian Newspapers. Journal of Peace Research(2), 6490. Gerhards, J. (2000). Die Entstehung einer europäischen Öffentlichkeit. In M. Bach (Ed.), Die Europäisierung nationaler Gesellschaften (pp. 277-305). Wiesbaden: Westdeutscher Verlag. Gerhards, J. (2002). Das Öffentlichkeitsdefizit der EU im Horizont normativer Öffentlichkeitstheorien. In H. Kaelble, M. Kirsch & A. Schmidt-Gernig (Eds.), Transnationale Öffentlichkeit und Identitäten im 20. Jahrhundert (pp. 135-158). Frankfurt am Main: Campus. Hagen, L. M., Berens, H., Zeh, R., & Leidner, D. (1998). Ländermerkmale als Nachrichtenfaktoren: Der Nachrichtenwert von Ländern und seine Determinanten in den Auslandsnachrichten von Zeitungen und Fernsehen aus 28 Ländern. In C. Holtz-Bacha, H. Scherer & N. Waldmann (Eds.), Wie die Medien die Welt erschaffen und wie die Menschen darin leben (pp. 59-82). Opladen: Westdeutscher Verlag. Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods. Riverside, CA: University of California, Riverside ( published in digital form at http://faculty.ucr.edu/~hanneman/ ). Harcup, T., & O'Neill, D. (2001). What Is News? Galtung and Ruge revisited. Journalism Studies, 2(2), 261-289. Harding, R. M., & Sokal, R. R. (1988). Classification of the European language families by genetic distance. Proceedings of the National Academy of Sciences USA, 85, 9370-9372. Hardy, J. (2008). Western Media Systems. London: Routledge. Ishii, K. (1996). Is the U.S. over-reported in the Japanese press? International Communication Gazette, 57(2), 135-144. Kepplinger, H. M. (1998). Der Nachrichtenwert von Nachrichtenfaktoren. In C. Holtz-Bacha, H. Scherer & N. Waldmann (Eds.), Wie die Medien die Welt erschaffen und wie die Menschen darin leben (pp. 19-38). Opladen: Westdeutscher Verlag. Kevin, D. (2003). Europe in the Media. A Comparison of Reporting, Representation and Rethoric in National Media Systems in Europe. London: Lawrence Erlbaum. Kim, K., & Barnett, G. A. (1996). The Determinants of International News Flow: A Network Analysis. Communication Research, 23, 323-352 Koopmans, R., & Erbe, J. (2004). Towards a European public sphere? Journal of Social Sciences, 17, 97-118. Koopmans, R., & Pfetsch, B. (2007). Towards a Europeanised Public Sphere? Comparing Political Actors and Media in Germany. In J. E. Fossum, P. Schlesinger & G. O. Kvaerk (Eds.), Public Sphere and Civil Society. Transformations of the European Union (pp. 57-88). Oslo: Arena. Kraus, P. A. (2004). Die Begründung demokratischer Politik in Europa. Zur Unterscheidung von Input- und Output-Legitimation bei Fritz W. Scharpf. Leviathan, 32(4), 558-567.

29

European Newsflows Meyer, C. O. (2005). The Europeanization of Media Discourse. A Study of Quality Press Coverage of Economic Policy Co-ordination since Amsterdam. . Journal of Common Market Studies, 43(1), 121-148. Machill, M., Beiler, M., & Fischer, C. (2006). Europe-Topics in Europe's Media: The Debate about the European Public Sphere: A Meta-Analysis of Media Content Analyses. European Journal of Communication, 21(1), 57-88. Meyer, C. O. (1999). Political Legitimacy and the Invisibility of Politics: Exploring the European Union's Communication Deficit. Journal of Common Market Studies, 37(4), 617-639. Nagpaul, P. S. (2003). Exploring a pseudo-regression model of transnational cooperation in science. Scientometrics, 56(3), 403-416. Nnaemeka, T., & Richstad, J. (1980). Structured Relations and Foreign News Flow in the Pacific Region. International Communication Gazette, 26, 235-257. Peter, J. (2003). Why European TV News Matters. A Cross-Nationally Comparative Analysis of TV News about the European Union and its Effects (Doctoral dissertation). Peter, J., & de Vreese, C. H. (2004). In Search of Europe: A Cross-National Comparative Study of the European Union in National Television News. Harvard International Journal of Press/Politics, 9(4), 3-24. Peter, J., Semetko, H. A., & de Vreese, C. H. (2003). EU Politics on Television News: A CrossNational Comparative Study. European Union Politics, 4(3), 305-327. Robinson, G. J., & Sparkes, V. M. (1976). International News in the Canadian and American Press: a Comparative News Flow Study. International Communication Gazette, 22(4), 203-218. Rosengren, K. E. (1970). International News: Intra and Extra Media Data. Acta Sociologica, 13, 96109. Rosengren, K. E., & Rikardsson, G. (1974). Middle East News in Sweden. International Communication Gazette, 20, 99-116.

Rosengren, K.E. & Windahl, S. (1989) Media Matter. TV Use in Childhood and Adolescence. Norwood, NJ: Ablex. Scharpf, F. W. (1999). Regieren in Europa. Effektiv und demokratisch? Frankfurt am Main: Campus. Schiller, H. (1976). Communication and Cultural Domination. Armonk, NY: Sharpe. Schulz, W. (1997). Changes of the mass media and the public sphere. Javnost - the Public, 4(2), 5771. Semetko, H. A., & Valkenburg, P. M. (2000). Framing European politics: a content analysis of press and television news. The Journal of Communication, 50(2), 93-109. Skurnik, W. A. E. (1981). Foreign News Coverage in Six African Newspapers: the Potency of National Interests. International Communication Gazette, 28(2), 117-130. Trenz, H.-J. (2000). Korruption und politischer Skandal in der EU. Auf dem Weg zu einer eurpäischen politschen Öffentlichkeit. In M. Bach (Ed.), Die Europäisierung nationaler Gesellschaften (pp. 332-362). Wiesbaden: Westdeutscher Verlag. Trenz, H.-J. (2004). Media Coverage on European Governance. Exploring the European Public Sphere in National Quality Newspapers. European Journal of Communication, 19(3), 291-319. Van De Steeg, M. (2006). Does a public sphere exist in the European Union? An analysis of the content of the debate on the Haider case. European Journal of Political Research, 45, 609634. Wallis, R., & Baran, S. J. (1990). The Known World of Broadcast News : International News and the Electronic Media. London Routledge. Wanta, W., Golan, G., & Lee, C. (2004). News: Media Influence on Public Perceptions of Foreign Nations. Journalism and Mass Communication Quarterly, 81(2), 364-377. Westerstahl, J., & Johansson, F. (1994). Foreign News: News Values and Ideologies. European Journal of Communication, 9(1), 71-89. White, D. M. (1950). The "gate keeper": A case study in the selection of news. Journalism Quarterly, 27, 383-390.

30

European Newsflows Wilke, J. (1998). Konstanten und Veränderungen der Auslandsberichterstattung. In C. Holtz-Bacha, H. Scherer & N. Waldmann (Eds.), Wie die Medien die Welt erschaffen und wie die Menschen darin leben (pp. 39-58). Opladen: Westdeutscher Verlag. Wu, H. D. (2000). Systemic determinants of international news coverage: a comparison of 38 countries. Journal of Communication, 50(2), 110-130. Wu, H. D. (2003). Homogeneity around the world? Comparing the Systemic Determinants of International News Flow between Developed and Developing Countries. International Communication Gazette, 65(9), 9-24.

i

The study is part of the Piredeu project that integrates various studies regarding the European Parliament Elections 2009 http://www.piredeu.eu/. ii The frontpage, a randomly selected page and all stories concerning the EU. iii As of 8.10.2009. The data collection is still in progress in two countries. iv According to a keyword search (“European Public Sphere”) in Google Scholar, conducted 4.10.2009 v In this paper the term public sphere refers to a ‘media constructed’ public sphere. Following the argumentation of Schulz (1997) this implies that the mass media are regarded as the constitutive element of a public sphere of a political community. vi Others, like the characteristics of the event covered cannot be considered in the present study, to avoid too much complexity. vii To organize this cooperation they formed a regional intergovernmental organization, the Nordic Council, after the Second World War. viii All stories that mentioned the EU(or synonyms) at least twice. ix 'The outlets selected are: Austria: ZiB 19.30 (ORF1); Aktuell 19.20 (ATV), Der Standard, Die Presse, Neue Kronen Zeitung; Belgium Het Journaal 19.00 (VRT),'VTM-Nieuws 19.00 (VTM), JT Meteo 19.30 (La Une), Le Journal 19.00 (RTL-TV)', De Morgen, De Standard, Het Laatste Nieuws', La Derniere Heure, La Libre Belgique, Le Soir; Bulgaria: 24 Chasa, Dnevnik, Trud, bTV 19.00, BNT kanal 1 20:00; Cyprus: RIK1 20.00, Ant1 20.15, Fileleytheros, Haravgi, Simerini; Czech Republic: Udalosti 19.00, Televizni noviny 19.30, Blesk, Mlada Fronta, Pravo; Denmark: Nyhederne 19.00 (TV2), TV-avisen 21.00 (DR 1), Ekstra Bladet, Morgenavisen Jyllandsposten, Politiken; Estonia: Aktuaalne kaamera 21.00 ( ETV), Reporter 19.00 (Kanal2), Eesti Ekspress Wochenblatt, Postimees, SL Öhtuleht, Finland: Tv-uutiset ja sää 20.30 (YLE TV1), Kymmenen uutiset 22:00 (MTV3), Aamulehti, Helsingin Sanomat, Iltasanomat, France: Le Journal 20.00 (TF1), Le Journal 20.00 (F2), Le Figaro, Le Monde, Libération; Germany: Tagesschau 20.00 (ARD), Heute 19.00 (ZDF), RTL aktuell 18.45 (RTL), 18.30 (SAT1), Bild, FAZ, SZ; Greece: 20.00 Mega, Eleftherotypia, Kathimerini, Ta Nea; Hungary: Hírádo 20:30 (M2), Esti Híradó 18:30 (RTL Klub), Blikk, Magyar Nemzet, Nepszabadsag; Ireland: Nine News 21.00 (RTEI1), TV3 News 17:30 (TV3), Irish Independent, The Irish Times, The (Daily) Star; Italy: TG1 20.00 (RaiUno), TG5 20.00 (Canale5), Il Corriere della Sera, Il Giornale, La Repubblica; Latvia: Panoramas 20:30 (LTV), T Zinas 20:00 (LNT), Diena, Latvijas Avize, Vesti segodnya, Lithuania: Panorama 20.30 (LTV), TV3 žinios 18.45 (TV3), Lietuvos rytas, Respublika, Vakaro zinios; Luxembourg: De Journal 19.30 (RTL), Tageblatt, Voix du Luxembourg, Wort (D), Malta: L-Abarijiet TVM 20.00 (TVM), One News 19.30 (One TV), Nazzjon, Orizzont, The Times (engl.); Netherlands: RTL Nieuws 19.30 (RTL), NOS Journaal 20.00, De Telegraaf, De Volkskrant, NRC Handelsblad; Poland: Fakt, Gazeta Wyborcza, Rzeczpospolita, Wiadomoci 19:30 (TVP1), FAKTY 19:00 (TVN); Portugal: Telejornal 20:00 (RTP1), Jornal Nacional (20:00) (TVI), Correio da Manha, Jornal de Notícias, Publico, Romania: Evenimentul Zilei, Jurnalul National, Libertatea, Telejurnal 20:00(TVR1), Stirile 19.00 (Pro TV), Slovakia: Spravy 19:30 (STV 1), Televizne Noviny 19:00 (TV Markiza), Daily Pravda, Nový cas, Sme/Práca; Slovenia: Dnevnik 19.00 (TV S1), 24UR 19.00 (POP TV), Dnevnik, Slovenske Novice, The Delo; Spain: Telediario-2 21.00 (TVE1), Telecinco 20.30 (Tele5), Noticias2 21.00 (Antena3), ABC, El Mundo, El Pais; Sweden: Rapport 19.30 (TV2), Nyheterna 18.25 (TV4), Aftonbladet, Dagens Nyheter, Svenska Dagbladet; UK: BBC1 News at 10, ITV News at 10, Daily Telegraph (Sunday: Sunday Telegraph), Guardian (Sunday: The Observer), Sun x Referenda draw a much higher media attention than regular elections though. xi The coders were asked to assign each news item to a particular country or geographic entity by answering the question: “Where does the story or the actions it depicts (mainly) take place (in terms of prominence in the story or length)?”, only those stories that concerned other European member states were included in the analysis (N: 8537)

31

European Newsflows

xii Added monetary value of traded goods and services as reported by the importing country in 2008. Extracted from Eurostat 1.10.2009 xiii Operationalized as the distance between the capitals. xiv According to membership in the Warsaw pact. xv Assigned by applying the typology provided by Harding and Sokal (1988). xvi As reported by the International Monetary Fund, World Economic Outlook Database, October 2008 edition xvii Extracted from Eurostat 1.10.2009 xviii Members of parliament after the Elections 2009 per member state. As reported on the homepage of the European parliament: www.europarl.europa.eu/members/expert/groupAndCountry.do?language=EN xix Contributions to the budget in 2008 as reported by the Commission: http://ec.europa.eu/budget/documents/2008_en.htm?go=t3_2#Table-3_2 xx For the purpose of visualization the data is symmetricized by calculating the average of newsflows from A to B and B to A. Symmetric data is a necessary condition for visualizations on a metric level.

32