Covering the Euro crisis: A content analysis on
Dutch, British and Italian newspapers.
Master Thesis Communication Science
Author: Roos van Odijk
Student ID: 5862523
Supervisor: Linda Bos
Date of completion: 30/01/2015
Graduate School of Communication
2 Abstract:
This study investigates the way the Euro crisis was covered in Dutch, British and Italian newspapers. In study 1/, the occurrence of six news frames in news stories about the Euro crisis was investigated. Furthermore, the tone of the actor in the news towards the ability to solve the Euro crisis was analyzed and the actors in the news were coded. A manual content analysis was performed on articles (n=300) in six Dutch, British and Italian newspapers. The prevalence of news frames around the two Greek bailouts in 2010 and 2012 was
analyzed and compared. Significant differences were found between the framing of the three countries. The results showed that, overall the economic consequences frame was most commonly used in the news about the Euro crisis. In study 2, the background of the actors in the news was coded, by performing an automatic content analysis. Three Python scripts were created to count the actors in news articles (n=960) in four periods in Dutch, British and Italian news. EU- actors, domestic government actors and political actors from other European countries were most present in the news articles. The results of this study imply that a country’s political and media system has consequences for the way news about the Euro crisis is presented to the public.
3
For the past three years, Europe has been mesmerized by the Euro crisis, namely the struggle to resolve the debt problems facing Greece, Ireland, Italy, Portugal and Spain without breaking up the monetary union or precipitating a wider financial crisis in Europe”
(Hall, 2012, p. 355). The Euro has been greatly affected by the global economic crisis that emerged in the summer of 2007 (Arestis & Sawyer, 2012). Several European countries have serious economic problems due to the economic crisis. Greece, however, has been hit the hardest from all the Eurozone countries. This ‘Greek tragedy’ made the Eurozone countries loan big sums of money to the Greek government in order to save the Greek economy. The Eurozone countries wanted to prevent Greece from leaving the Euro currency and return to the Drachma. This scenario is also known as the possible ‘Grexit’. Not just the Greek economy itself was at stake. Also fear of the contagion of the severe financial situation of Greece on the other Eurozone countries arose (Arestis & Sawyer, 2012). Besides this, the image of the Euro in the world and on the financial markets had to be rescued. This image is important in the context of the legitimacy of the European integration project according to Trenz (2012). The Euro crisis has brought a new dimension in the way citizens, media and political elites perceive the legitimacy of the EU and the Euro. The economic crisis caused a deeper divide between different countries, left- and right and pro- and anti- European camps. Important is the influence of the media in this process. This is because the opposition and support with the European integration project is expressed through the media (Trenz, 2012). Research has shown that most European citizens receive their information about European affairs through the media (Eurobarometer, 2012). By framing European issues in different ways, the media can alter public opinion and cause a shift in policy (Terkildsen & Schnell, 1997). Moreover, the formation of public opinion can have important consequences for the outcome of European elections (Radu & Stefania, 2012). Therefore it is important to do research on the occurrence of different frames in the media (Valkenburg et al., 1999).
4 This research consists of two studies that both investigate the way European media covered the Euro crisis. The investigated European media outlets in this study are newspapers from three different European countries, namely the Netherlands, Great Britain and Italy. Articles from six newspapers, both left- as right-wing, will be analyzed. These countries all represent a different model that characterizes different media and political systems (Hallin & Mancini, 2004).
The first part of this research, study 1, withholds a deductive approach that involves the analysis of predefined, generic frames with the aim to find out the extent to which these frames are present in the news (Semetko & Valkenburg, 2000). According to de Vreese (2001), little attention has been devoted to the framing in a cross-national comparative way. This research therefore contributes to the literature about framing in a cross- national comparative situation. Moreover, Schuck and de Vreese (2006) state that it is striking that, because of the increased political relevance of the EU, studies that deal with framing of issues in this context are still very limited. Therefore this study also tries to contribute to the closure of the gap in the literature about the framing of European affairs.
Besides this, in study 1, the tone of the actor in the news towards the ability of the involved actors to solve the Euro crisis will be analyzed. There have been concerns that the portrayal of the EU in the news has a negative and Eurosceptic tone (Norris, 2000). This negative tone could cause a growing cleavage between the European leaders and its public. This illustrates why it is important to look at the tone, in which news about European affairs, in this case the Euro crisis, is presented.
Finally, the background of the actors present in the selected news articles will be analyzed in both studies. According to Gleisner and de Vreese (2013), this has to do with the Europeanness of a country, the extent to which the focus of the news is either on national or on EU actors and issues. First the analysis of actors in the news will be done manually in
5 study 1. Secondly, a Python script will be created to count the actors automatically in a second study . The aim is firstly to find out whether there are differences in the use of actors in countries and over time. However, besides this, the aim is to discuss the implications of the different methods. How do a manual and an automatic content analysis differ in this study?
The research question that I pose, with the aim to distinguish differences in the use of frames, tone and actors between countries with different media and political systems,
concerning the issue of the Euro crisis, is as follows:
To what extent do Dutch, British and Italian newspapers differ in their coverage of the Euro crisis?
News frames
Many researchers have devoted attention to the theory of framing, among them Entman (1993), who states that: “Framing is selecting some aspects of a perceived reality to
enhance their salience in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation” (p. 53). The media not
just present the reality, but define, create and shape a reality that is seen as a true one (d’Haenens, 2005). Framing can make a message more salient, which means making a message more evident, meaningful or memorable to the public (Entman, 1993). An increase in salience makes it more likely that an audience will perceive the information given, recognize the meaning and thus process it and place it in their memory.
News frames are powerful because they decide the spectrum in which citizens discuss public events (Tuchman, 1978). This is confirmed by Neuman et al. (1992), who states that: “News frames can be defined as conceptual tools which media and individuals rely on to
convey, interpret and evaluate information” (p. 60). Moreover according to Entman (1993),
6 power. This is because the frame in a news message identifies the actors or interests that compete with each other and with journalists to dominate the news frames.
Issue specific frames and generic frames
When investigating the occurrence of these news frames in a news story, a distinction can be made between issue specific frames and generic frames (De Vreese, Peter & Semetko, 2010). Issue specific frames relate to specific topics or news events. They can be applied only in some contexts. If an issue specific approach is chosen, there is more room for the investigation of the framing of an event or issue in detail.
However, this approach makes it more difficult to generalize and make comparisons between frames, topics and framing in different countries. Studies of issue specific frames investigate the occurrence of frames that are linked to the particular issue that is in question (De Vreese, 2005). This opposed to generic news frames, which are frames that are structural and are linked to the conventions of journalists. Semetko and Valkenburg (2000) identified five different generic frames. The first one is the Conflict frame, which emphasizes conflict between individuals, groups or institutions as a means of capturing audience interest (Neuman et al., 1992, p. 61-62). It emphasizes the points of disagreements between those parties in conflict (De Vreese, 2001). The conflict frame is a commonly used frame in news about EU related topics (Gleissner & De Vreese, 2005). The conflict frame is linked to strategy coverage, in which winning and losing is central and the competition is a main feature (Jamieson, 1992). Besides this frame, Semetko and Valkenburg (2000), also identified the Human interest frame which gives a human face or shows the emotional part of an issue, event or problem to the public. By using this frame the aim is to bring some drama to the story and to personalize the news. By personalizing the news the character of the news is becoming more narrative. Another frame that is commonly used in the news is the Economic
7
consequences frame that emphasizes a problem or issue in the context of its economic
consequences for a group, a country or an individual. According to Gamson (1992),
journalists often frame the news in terms of economic consequences with the aim to make a certain issue relevant to the audience. In a study performed by Semetko and Valkenburg (2000), the economic consequences frame appeared to be the most common frame in the news. As opposed to the economic consequences frame, the attribution of responsibility
frame presents the news about an issue, event or problem not in terms of economic
consequences but in a way that responsibility for its cause or solution is attributed to a person a group or a government. According to Semetko and Valkenburg (2000), the
attribution of responsibility frame was mostly used in serious and sober newspapers. In their study these newspapers are outlets that present more political and economic news. Less serious newspapers on the other hand, are more sensationalist, are aimed at a broader public
and much easier to read. The fifth generic frame is the Morality frame, which focuses on a problem or issue by
highlighting the religious or moral context (Semetko & Valkenburg, 2000). Journalists are supposed to stick to the professional norm of objectivity. However, they often refer to moral frames indirectly by letting someone else raise a question, through inference or by quoting that person (Neuman et al., 1992). A news outlet could, for example, use the opinions of an interest group to raise questions about gay marriage or abortion.
Besides the five generic frames, the presence of one issue specific frame will be analyzed in this research, namely the powerlessness frame. ‘This frame is based on the idea that the relationship among key actors of a story is unbalanced, unfair and thus, the need to identify the ones to blame for the current situation is rather urgent (De Vreese, 2005, p. 56)”. Neuman et al. (1992), identified the powerlessness frame in their analyses of the media coverage of some current issues. This frame referred to ‘the dominance of forces
8 over weak individuals or groups (Neuman et al., 1992, p. 67). Radu and Stefanita (2012) used this frame in their analyses of the framing of the Economic crisis in the Romanian online media.
The Netherlands, Great Britain and Italy representing three models of media and
politics
The occurrence of frames in newspapers of three European countries, the Netherlands, Great Britain and Italy will be investigated. These countries all represent one of the three models that were distinguished by Hallin and Mancini (2004). The models characterize different media and political systems in the western world. The first one is the Liberal or North Atlantic model, which prevails around North America, Canada, Ireland and Great Britain. The second one is the Democratic corporatist or North/Central
European model, that exists in countries like Austria, the Netherlands, Germany, the
Scandinavian countries and Belgium. Finally, the polarized pluralist or Mediterranean model that prevails in southern Europe, in France, Greece, Italy, Spain and Portugal. These models are based on four dimensions (Hallin & Mancini, 2004).
The first dimension, the development of media markets deals with the question how developed the mass circulation press in a media system is. This is about quantity, thus how many newspaper sales per population. The circulation of newspapers is the highest in the democratic corporatist countries, followed by the countries belonging to the liberal model and finally by the polarized pluralist countries. The development of media is also about the relation with its audience and its part in a bigger process of social and political communication. A horizontal process of communication means that a small elite is addressed by newspapers. Their content refers to well-educated and politically active citizens. The polarized pluralist model is known for its horizontal process of communication. Italy is a
9 good example in this context. The Italian media focus on the high-educated and politically engaged public. Tabloid or sensationalist papers do almost not exist in Italy and the lower-educated citizens make more use of electronic media. In a vertical process of communication on the other hand, newspapers mediate between political elites and the ordinary citizen. In these systems newspapers address their news content to the mass and not necessarily to the politically engaged and high educated citizens. Great Britain, that represents the liberal model, has a class-stratified newspaper market. This means that there is a sharp divide between the quality newspapers, that are aimed at the middle- to upper-class reader and the tabloids which are more sensationalistic and aimed at a broad public. The Democratic corporatist countries are characterized by a wide range of different newspapers.
Next to the development of media markets, is the degree of political parallelism in a media system also a dimension that differs between the systems. The degree of political parallelism explains the extent to which the structure of the media system reflects that of political tendencies in a country. When focusing on media content this means whether media reflect political orientations in their news content. In the Netherlands, just like in other countries representing the democratic corporatist model, the newspapers have a distinctive political orientation. An example is the newspaper Trouw, which is progressive in character and is oriented towards the third world and the environment. In Great Britain, the newspapers became less supportive for one party over time. Despite this decline of political parallelism, the British newspapers have distinct political orientations. This is however especially the case with the tabloid media, usually a right-wing, populist orientation. According to Hallin & Mancini (2004), most Italian newspapers, even the commercial newspapers, are linked to a political party. Besides the reflection of political tendencies in the media, the role of a journalist is important when deciding the degree of political parallelism (Hallin & Mancini, 2004). The journalistic culture in a country is an indicator of political parallelism. In a
10 country with strong political parallelism, the reporting style of journalists is closely related to that of politics. In these countries the journalist moves more freely between the role of commentator and reporter, while in other countries a journalist sticks to the role of reporting the news. The democratic corporate countries have a strict separation between commentary and reporting Dutch journalists are known for their objective way of writing. In Great Britain, however, especially in election periods, the journalists writing for tabloid media have a high degree of partisanship, by supporting a campaign of a political party. The quality papers have a more subtle style of writing. In Italy and other countries belonging to the polarized pluralist model, a commentary-oriented style of journalism is common. Journalists comment on the presented issues in the news article.
Also dealing with journalism is the third dimension, journalistic professionalization. A key part of professionalization is the journalistic autonomy which varies heavily between media systems and media outlets. The degree of professionalization is determined by shared norms like ethical principles. Besides this, the journalistic
professionalization depends on the instrumentalization of the media. This is the control of actors outside the media outlet like politicians, economic actors or parties. Especially the liberal model is known for its strong professionalization of journalism. In Great Britain journalists have their own set of criteria for presenting a story. The democratic corporatist model has a lower professionalization of journalism
. The journalistic autonomy in a Mediterranean country is low. This is especially the case in Italy, where the players in Italian politics have a significant control over the newspapers in Italy. A saying in Italy is that a journalist is known as a ‘giornalista
dimezzato’, which literally means a journalist split in halves. One half belongs to the
journalist himself, the other half belongs to actors such as media owners, politicians and investors (Hallin & Mancini, 2004).
11 Finally, the role of the state decides to which model a Western country belongs. Public service broadcasting is the most important form of state intervention. Besides this there are other forms of state intervention such as state owned news agencies and newspapers or for example hate speech or right-of-reply laws. In the countries of the liberal model, the role of the market is bigger than that of the state. This is caused by the strong liberal tradition in these North Atlantic countries. The Dutch system is known for its high level of external pluralism, with separate broadcasting companies representing different social groups and religions. Despite of a high the state intervention in the democratic corporatist countries, the independence of journalists is also highly protected. In Southern Europe, the state plays a big role in the media market. Business has a big influence in state contracts, subsidies and regulations. Therefore business owners have been willing to subsidize media outlets in these countries (Hallin & Mancini, 2004).
Implications of the Dutch, British and Italian media systems on the differences in framing
The characteristics that belong to the different media systems of the three countries have implications for the choices journalists make, for the way the issue is presented to the public, thus for the frames that are used in an article (Van Dalen et al., 2011). In different studies, researchers have looked at the use of generic frames in the news.
Firstly, the human interest frame is mostly used in a liberal news market like the British one (Semetko & Valkenburg, 2000). This is linked to the character of such a liberal market, which is very competitive and aims to answer to the demand of the public. The human interest frame contains the dramatization and personalization of the news. Moreover they feature more often the private lives of individuals in news stories. All of this makes the news more entertaining. This entertainment has the goal to give the public what it wants.
12 Also van Daalen, De Vreese and Albaek (2011), confirm this after conducting a survey among political journalists in four European countries, including the United Kingdom. This survey showed that British journalists are more entertainment oriented than Danish, German and Spanish journalists. According to Semetko and Mandelli (1997), the aim of increasing the audience share goes along with the use of the human interest frame, with a focus on personalities and individuals. Furthermore, a research about the framing around the
introduction of the Euro, showed that the human interest frame was barely used in the Dutch news (De Vreese et al. 2011). Besides this, according to d’Haenens (2005), who identified news frames in news about EU affairs, the human interest frame was almost non-existent in the Spanish news. According to Hallin & Mancini (2004), Italy and Spain belong to the same model of Mediterranean countries and therefore the same expectations can be made. The first hypothesis concerning the human interest frame is as follows:
H1: The human interest frame will be mostly used in the British newspapers
When focusing on the conflict frame, some similar assumptions can be made as with the human interest frame. According to van Dalen (2012), the use of a conflict frame has, just like the human interest frame, the goal of entertaining the audience. In a news market that is as competitive as the British one, journalists and editors have to do everything to get the attention from the public (Bennett, 1995). One way of achieving this is by the use of the conflict frame (Semetko & Mandelli, 1997). Firmstone (2011) investigated the way the British press frames the news around the Euro. She found out that the British press framed the news around the Euro in terms of conflict, for example about the lack of consensus in the elite opinion. De Vreese et al. (2000), investigated the occurrence of frames in print and television news in news about Europe in Dutch, Danish, British and German newspapers. They found out that the conflict frame was the dominating frame in economic and political news. Besides this they concluded that in general, in the news that was
13 particularly about the Euro introduction, the economic consequences frame was more present than the conflict frame. Striking was however that only in the British news the news around the introduction of the Euro was significantly more framed in terms of conflict than in terms of economic consequences. Based on the theory described above, the second hypothesis concerning the use of the conflict frame in the Dutch, British and Italian newspapers, is as follows:
H2: The conflict frame will be mostly used in the British newspapers
The economic consequences frame is used to make an issue relevant to the public (Gamson, 1992). Moreover, by using this frame, the event has a big chance of becoming a news story (Graber, 1993). This is because consequence and the wider impact of an event are considered to be central news values. De Vreese (2012), states that there is a relative absence of the economic consequences frame in the British news media. This may be explained by the findings of Kriesi (2012), who states that Dutch journalists are willing to interpret and add analyses to their news. This compared to British journalists who consider themselves merely as transmitters of facts instead of interpreters of news. Besides this, de Vreese et al. (2001), found out that in stories about the introduction of the Euro, Dutch journalists make the most use of the economic consequences frame compared to other frames. Also, compared to the British journalists, they framed the stories about the Euro more in terms of economic consequences. The Dutch news focused for example on the conversion work in the financial sector and the consequences of the Euro for the Dutch economy. Their British counterparts however, did not devote a lot of attention to important economic repercussions for the British pound (De Vreese et al., 2001). Based on the outcomes of the previously mentioned studies, the hypothesis concerning the use of the economic
14
H3: The economic consequences frame will be mostly used in the Dutch newspapers
The morality frame is according to Neuman et al. (1999) not very common in the news. However the Italian journalists might use this frame more than Dutch or British journalists. This is because journalists in Italy have a more commentary- oriented style of journalism (Hallin & Mancini, 2004). There is a tendency to frame stories as moral scandals in the Italian media. Moreover, Italian newspapers often mobilize their readers by taking a political standing (Hallin & Mancini, 2004). Besides this, d’Haenens (2005) found out that in news about EU affairs, the morality frame is positively present in the Italian media. Triandafyllidou (2003), investigated how the launch of the Euro was covered in the Italian media. She found out that the historical and symbolic relevance of the Euro and Italy’s unity with Europe was emphasized. The launch of the Euro was framed as being part of the national pride for the Italians in different media outlets. This illustrates the
commentary- oriented style of journalism. According to Neuman et al. (1993), journalists do not use the morality frame that often because of the rule of objectivity. However, Italian journalists are known to be less objective than journalists from countries from the liberal model the polarized pluralist model (Hallin & Mancini, 2004). The hypothesis concerning the use of the morality frame in articles that present news about the Euro crisis is thus as follows:
H4: The morality frame will be mostly used in the Italian newspapers
In countries with strong social welfare systems ,like the Netherlands, the attribution of responsibility frame is commonly used because the news media attributes the responsibility for social problems to the government (Semetko & Valkenburg, 2000). In by Semetko & Valkenburg’s study (2000), on print and television news in the Dutch media, the attribution of responsibility frame was most commonly used. This is confirmed by Dirikx and Gelders (2010), who investigated the occurrence of generic
15 frames in the news about the climate crisis in both Dutch as French newspapers. They found out that the Dutch newspapers, especially the newspaper NRC Handelsblad, often makse use of the attribution of responsibility frame when presenting news about the climate crisis. In these articles the need for urgent action was mentioned, solutions for the problems were proposed and it was suggested that governments were responsible for the climate crisis or were able to alleviate the climate crisis. Therefore the hypothesis concerning the attribution of responsibility frame is as follows:
H5: The attribution of responsibility frame will prevail in the Dutch newspapers
The Issue-specific frame, the powerlessness frame, has not been investigated in studies as much as the generic ones. This is because it applies especially to a certain issue, in this case the Euro crisis (Radu & Stefania (2012). Firstly, according to Firmstone (2012), it is common in the British media when it comes to news about the Euro, to express the inability of actors concerning the introduction of the Euro. This is because there are many conflicts about the Euro and the contribution of the United Kingdom around the currency. Firmstone (2012), investigated the way the British newspapers presented the news about the introduction of the Euro. She found out that this did not express hope that the Euro would be a good investment. When focusing on the situation of Great Britain in the EU, it is interesting in this case, that Great Britain is known as a Eurosceptic, nonmember of the monetary union (Semino, 2002). Therefore you could argue that they are more skeptic towards the ability of the Eurozone countries to solve the Euro crisis and more critical towards the balance of power between the European countries. The hypothesis about the powerlessness frame in this study is as follows:
16 The tone of the actor in the news
Besides the use of the described frames, in this study the tone of the actors in the news towards the EU and the ability to solve the Euro crisis will be analyzed. According to Neuman et al. (1992) journalists can express their opinion about an issue by quoting someone in an article and letting an actor raise questions. By using the opinion of a certain group or person, while leaving out the opinion of other groups or persons, the journalist can take a stand on a certain topic. According to de Vreese and Gleissner (2005), the tone of the news coverage of political actors in the news is important for the evaluation of these actors given by the public. They performed a content analysis on television news and national newspapers and investigated how these media outlets covered the news about the Convention’s preparation of the European Union constitution. To assess the tone of the article, they looked at the stance of the actor towards the European Union constitution, in Dutch, British and German media. They found out that the German journalists reported slightly positive towards the Constitution. The Dutch journalists, followed by their British counterparts, reported in a negative way towards the constitution. Gleissner and De Vreese (2005), state that there is a love - hate relationship between the British media on one side and European affairs on the other side. According to d’Haenens (2005), the Netherlands is a country which is in general positive towards the EU. Gleissner and de Vreese (2005) state that this does not mean that the tone of the news is also positive. The tone in the Dutch media on European affairs tends to be rather mixed. In another study by de Vreese (2002), the news in both the British as the Dutch news about several European affairs, such as European elections and the introduction of the Euro, also had a negative tone of coverage. According to Triandafyllidou (2003), Italy is seen as a country that is very positive towards the EU and the Euro, which is reflected in the media content. You could however question whether the Euro crisis and the economic consequences of the Euro crisis for Italy have influenced this tone
17 (Arestis & Sawyer, 2012). In order to distinguish the differences between the tone in the news, the following research question is posed:
RQ1: To what extent does the tone of the actors in the news towards the ability of the involved actors to solve the Euro crisis differ between the three countries?
The background of actors in the news
According to de Vreese (2003), not only framing, but also the visibility and depiction of actors in the news may play an important role in the shaping of public opinion on Europe. Visibility of actors in the news is important because it decides the way the public looks at certain political issues.
Several studies have been conducted on the use of different (political) actors in foreign news. According to some researchers, the focus of a news story and thus also the use of actors in this story, has to do with the theory of proximity. The focus of a study and the use of actors depends on the proximity of the topic or the actor to the news outlet. According to Shuck et al. (2010), the closer to Brussels a European country is, the more the focus lies on EU news. The further away, the more the media of a country makes use of domestic actors. This thus depends on the geographical proximity. Also Martin (2007) looked at geographical proximity. He states that the variety and level of actors in a story depends on the choices of a journalist, which is an issue of news bias. Journalists are inclined to use actors in a story that are geographically close. According to Hackett (1984), who is well-known for his study on proximity, this is because of the unfamiliarity of journalists with actors other than those geographically close to them. Therefore it could be argued that newspapers make more use of domestic actors than international or European actors in their news stories.
18 According to Gleissner & De Vreese (2013), not many studies have been conducted on the particular visibility of EU actors in different media outlets. They state that the presence of these EU actors in the news compared to other (national) actors is important for the degree of Europeanness of a country (Gleissner & De Vreese, 2013). The Europeanness of a country deals with the extent to which the news in a country is domesticized, thus whether the focus of the news is either on national or on EU actors and issues. The assessment of Europeanness focuses on the attention of national media towards EU actors and organizations. The
prominence of EU actors versus domestic actors in the news can thus provide evidence for the extent of Europeanness of the news. This is because the presence of political actors on a European level has shown to be an important condition for the functioning of political representation in a EU democracy (De Vreese, 2002). When covering European news, the larger the focus of the news media in a country on European actors, the more those news media present it as a European event instead of a national event. Studies have shown a very diverse use of domestic versus EU actors in different EU countries. According to de Vreese (2003), a combination of the level of support for the EU and the differences in media and political systems between countries, decide the extent of EU or domestic in the news. This is supported by Featherstone (2003), who states that the national media of a country, with their media output, are responsible for the level of Europeanisation in a country. This level can be found by analyzing the media’s focus on national versus European issues and actors.
Schuck et al. (2011) looked at the extent of such Europeanness in the coverage on the elections in the different member states. They analyzed the news coverage of the 2009 EP elections in all the 27 member states. Their main conclusion was that even when EU issues were the subject of the news story, domestic political actors still dominated the story. De Vreese et al. (2006), on the other hand, found some differences between countries after he studied the prominence of EU versus domestic actors in the news around the 2004 European
19 parliamentary elections in EU-countries. They found out that national domestic actors
dominated the news about the European parliamentary elections in all 15 countries.
It is difficult to formulate hypotheses about the use of actors based on previous studies. This is because over a longer period, the level of Europeanization both inside a country as inside a newsroom can change. Brüggeman and Kleinen-von-Köningslöw (2009) analyzed the level of Europeanization in different European newspapers and found out that newspapers have different views on European issues. For example, the French newspaper Le Monde, was found to be a very Europeanized newspaper in comparison with the English newspaper The Times. They state that this could be because of the editorial mission of Le Monde, who wants to cover many European issues and actors. On the other hand this difference could also be explained by the disinterest of the British public in European actors and issues. We cannot just expect from a country’s political agenda towards Europe how high the level of Europeanization of the media in a country will be (Trenz, 2004). Also a
newspapers’ agenda and different news values will decide how high this level will be and thus in the context of this study the use of different actors in news stories.
The previous studies all investigate the use of domestic versus European actors. Most of this research has been done around European elections. However, it could be expected that news about the Euro crisis also contains many Greek politicians, leaders of European
countries and other international leaders. Therefore, in this study, the aim is to investigate the occurrence of EU-actors and domestic actors but also Greek actors, international actors and actors from EU countries, such as Angela Merkel or Nicolas Sarkozy. Besides this, it would be interesting to find out whether this use of actors has changed over time. Did for example increase the role of international actors in the news about the Euro crisis? To find out how
20 visible different actors are in the Italian, Dutch and British newspapers and whether this has changed over time, the following research questions are posed:
RQ1: How visible are European actors in Dutch, British and Italian news stories compared to other actors?
RQ2: To what extent did the visibility of actors in Dutch, British and Italian news increase or decrease from March 10, 2008 until February 8, 2014?
Methodology
Study 1
A content analysis is performed, using pre-defined frames. In order to measure frames concerning the Euro crisis, five generic frames and one issue-specific frame are added to the analysis. The aim is to measure the generic frames identified by De Vreese (2005), the
conflict frame, the human interest frame, the economic consequences frame, the attribution of responsibility frame and the morality frame. One issue-specific frame was added to the
analysis, namely the powerlessness frame.
The questions that Radu and Stefania (2012) used to measure the powerlessness frame in their research were adapted to the current research concerning the framing of the Euro crisis. Each frame contains binary questions that can be answered by either yes (1), or no (0). In order to make a comparison between the countries and newspapers on the tone towards the Euro crisis, the actors in the news will be identified. Moreover, for all actors identified, the tone towards the ability of solving the Euro crisis will be coded. “An actor is defined as an individual (or organization/ institution/entity etc.) who is an essential part of the content of the story and who is mentioned or shown in the news story” (De Vreese, 2003, p.86).
The background of the actor will also be coded, using the categories of de Vreese (2003). Firstly EU actors, which are (members of) EU institutions and parliamentarians. This
21 is followed by Domestic government actors, such as ministers and members of government parties. The domestic political actors are members of the opposition parties in the
government. The category national institutions withhold for example courts and the police. The next category is organizations; these are actors like companies and trade unions. Non-organized actors are actors such as experts, victims, individuals or celebrities. The category international level actors withhold for example the US president or an actor from the Middle East. Finally there is a category ‘other’ for the actors that cannot be placed in one of the previous mentioned categories. After coding the articles, the category non-organized actors was divided into laypersons and experts and one extra category was added, namely the category traders and investors.
An appendix with a codebook, with examples in case the questions to measure the frames were answered with a ‘yes’, is added to this research.
Period of study
The news coverage around two events during the Euro crisis will be analysed. These are the first and the second Greek bailouts, which were implemented in order to save Greece financially and keep them in the Euro zone. The news about the first Greek bailout was announced on the 2nd of May 2010. The Euro countries and the IMF decided to support Greece with a loan of 110 milliard euro’s in order to stabilize the Greek financial markets. The second period of study is around the second Greek bailout which was on the 20th of February 2012. On the condition that Greece would further implement budget cuts, Greece was given another loan of 130 milliard euros. The news around these two bailouts will be analysed, from one week before until one week after the bailouts. These events are chosen because they mark important peaks in the economic crisis (Radu & Stefania, 2012)
22 Media sample
In total 300 articles from six Dutch, British and Italian newspapers
will be analyzed. Thus 100 articles per country, equally divided over left- and right wing newspapers and the two time periods. Firstly, in the databases of the six newspapers, the articles were found by searching for articles with the topic ‘Euro crisis’ one week before and one week after the two bailouts. After this, a random sample, per newspaper and period, was taken of all the search results in the databases. In every country two quality newspapers will be coded, a newspaper that is leaning towards a left- wing political orientation and a
newspaper that is leaning towards a right- wing political orientation. Choosing only quality newspapers allows for a cross- national comparison (Semino, 2002). Besides this, by analyzing newspapers on the center-left and newspapers on the center-right of the political spectrum, a comparison can be made between the framing of newspapers with different political orientations (a broad spectrum of opinions). In the Netherlands the newspaper De Volkskrant and de NRC will be analyzed, with the Volkskrant leaning towards a left- and the NRC towards a right- wing political orientation (Gleissner & De Vreese, 2005). In the UK the Guardian, a left-of-center newspaper, and the Daily Telegraph, a right- of center
newspaper, are selected (Hallin & Mancini, 2004). The Italian newspapers to be analyzed are La Repubblica, orientated towards the left and Il Corriere della Sera, orientated towards the right (Semino, 2002).
Development of scales to measure the news frames
A principal component analysis with varimax rotation with the 26 framing questions was conducted to find out whether they reflect underlying dimensions. Five items had factor loadings that were too low to include in the scales, namely lower than 0.50. These items were: does the story refer to winners or losers (N for ‘yes’ = 16), does the
23 story suggest that an individual or group is responsible for the issue-problem (N for ‘yes’= 43), does the story make reference to morality, God, and other religious tenets (N for ‘ yes’= 0), does the story suggest an imbalance of power/ status among EU countries (N for ‘yes’=84) and does the news story suggest that the citizens of countries that are in need of financial help (Greece) are victims/second rate citizens/discriminated (N for ‘yes’= 71). Chrohnbach’s alpha was used to measure the internal consistencies for the six scales. The alpha scores were 0.69 for the economic consequences frame, 0.75 for the conflict frame, 0.75 for the human interest frame, 0.57 for the attribution of responsibility frame, 0.56 for the morality frame and 0.65 for the powerlessness frame.
Study 2
Period of study
The news coverage around four time periods was analyzed. These time periods all cover a two week period of news coverage and two years lie between all four periods. Firstly, a period has been chosen from before the Euro crisis started, thus before the news that several European economies were in a problematic shape. This period is approximately two years before the first Greek bailout, thus before the second period that was analyzed. This period ranges from March 10 until March 24, 2008. This period was chosen because no election took place in this period, thus the economic news coverage about Europe could not be overshadowed by election news. Secondly, the coverage around the first and second Greek bailouts were analyzed, which were implemented in order to save Greece financially and keep them in the Euro zone. These are the same time periods as used in study 1. The second period ranges from April 25 until May 9, 2010 and the third period from February 13 until February 27, 2012. The fourth period that was chosen was two years after the second Greek bailout. On the first of February 2014, the Trojka announced that they considered to give
24 Greece another loan. This however depended on the approval of the EU member states. Because of the threat of another bailout, and thus a comparison with the first and second period, this fourth period has been chosen. The fourth period ranges from January 25 2014 until February 8, 2014.
Media sample
In total, 960 articles from six Dutch, British and Italian newspapers were analyzed. The 300 articles from study 1 were complemented with 660 more articles. Again, the
databases of the newspapers were searched for articles that had the topic ‘Euro crisis’. In the first period, also the search words ‘debt crisis’ and ‘financial crisis’ were used, because officially the Euro crisis did not start yet in 2008. After this, a random sample was taken from the databases.
Analytic strategy
Three python script were created in order to process and analyze all the newspaper articles, one for each of the three countries. According to Odijk et al. (2013), with a Python script, larger volumes of news articles can be analyzed. Moreover, the reliability of an automated content analysis is high because of its replicability. This replicability also allows studying long-term trends. De Fortuny et al. (2012), used a Python script in their study on political media coverage in Belgium. They state that an important advantage of a Python script is that an objective overview of the way of reporting can be given.
To investigate the occurrence of actors in the news, as in study 1, the actor categories of de Vreese (2003), have been used. The categories are: EU-actors, Domestic government actors, domestic political actors, other EU-country actors, Greek actors and International actors, the Greek actors have been added. To count the actors, different search strings were
25 created, withholding names of the most-used actors, such as Obama, Draghi, Merkel and Greek political actors (a list of actors is included as an appendix). Roggebrand and
Vliegenthart (2007) studied the occurrence of frames with a computer assisted coding. They created their search-strings by first including factors from their qualitative study. These were complemented with factors from other documents, such as parliamentary documents. In this study the same has been done by first creating search strings based on the articles that were manually coded in study 1. These were supplemented with more names from politicians. Important was that this list was exhaustive because in every period different actors were in the government or opposition parties. The complementary names were found on different internet sites, such as the website of the documentation center of the University of Leiden, that has an archive of all Dutch politicians. Criteria were decided for the list of actors. From the government, the prime-minister should be included, the minister of foreign affairs and the minister and secretary of state of finance. From opposition parties, the party leader should be added. Furthermore, the country leaders of the Eurozone countries were included. The list of international actors was based on the articles from study 1.
The article written by Trilling (2014), was used to create the Python scripts. Firstly, the amount of times an actor is counted in an article was counted and automatically divided into one of the six categories. The last name of the actor was included, because many times ony his or her last name is written. The articles were numbered so that the result was saved per article. If an actor was counted for example five times in an article, this was included in the results. After running the script once, fifty articles were compared to the Python script. Some names of actors were included in the script, after finding out that they were present in the article but not in the script. A comment was included to the Python scripts in order to produce CSV files, in which all the results were saved per category (Trilling, 2014). The CSV files could be imported to the statistical software SPSS, to perform analyses.
26 Results
Study 1
Table 1.
Percentage of Frames used in six newspapers in three countries
Netherlands Great Britain Italy
Frame Volks- krant NRC The Guardian The Daily Telegraph La Repubblica Il Corriere della sera Total % Economic Consequences 79% 55% 69% 63% 52% 59% 63% Conflict 24% 39% 67% 45% 39% 23% 39% Human Interest 10% 7% 30% 26% 14% 11% 16% Responsibility 45% 25% 30% 30% 44% 37% 35% Morality 4% 3% 3% 8% 11% 13% 7% Powerlessness 28% 20% 29% 37% 15% 15% 24% Occurrence of frames First of all, the news about the Euro crisis in the Dutch, British and Italian newspapers was mostly framed in terms of economic consequences (see table 1). In 63% of all the articles the economic consequences frame was found. This was followed by the conflict frame, which was present in 39% of all the analyzed articles and the attribution of responsibility frame,
27 present in 35% of the articles. The human interest frame was only used in 16 % of the
articles. Moreover, the morality frame was barely used, with a score of 7% in all the articles.
Visibility of news frames in Dutch, British and Italian newspapers
The six hypotheses concern the question whether a difference existed between the framing of the news in newspapers in the Netherlands, Great Britain and Italy. To investigate whether the use of frames varied between the countries, six-one way analyses of variance (ANOVA), were carried out, with the different types of news frames as dependent variables and the three countries as the independent variable. The ANOVA showed significant differences in the use of frames between the three countries.
First of all, a significant association was found between the countries concerning the use of the human interest frame, F(2.297) = 17.2, p < .001. A post-hoc test with Bonferroni indicated that there was a significant difference in the use of this frame between the British and the Dutch newspapers (Mdifference = 0.19, p < .001) and between the British and the Italian newspapers (Mdifference = 0.16, p < .001). No significant difference between Dutch and Italian newspapers concerning the use of the human interest frame was established. The human interest frame was mostly used in the British newspapers compared to the Italian and the Dutch newspapers (see table 4). The first expectation, that the human interest frame was mostly used in the British newspapers is therefore supported.
Concerning the conflict frame, a similar result was found. A significant association between the country and the use of the conflict frame was found, F( 2.297) = 13.6, p < .001). The post-hoc test showed that again the British newspapers made significantly more use of the conflict frame in comparison with the Dutch newspapers (Mdifference = 0,06, p < .001) and in comparison with the Italian newspapers (Mdifference = 0.06, p < .001). Besides this, no significant difference in the use of the conflict frame was found between Italy and the
28 Netherlands. The British newspapers made the most use of this frame in comparison with the Dutch and the Italian newspapers. Hypothesis 3, that the conflict frame would be mostly used in the British newspapers, can be confirmed.
Also the morality frame was significantly more used in the analyzed newspapers, F (2.29), = 4.6, p < .05). The post-hoc test revealed that there was a significant difference between the Italian and the Dutch newspapers (Mdifference = 0.09, p < .05). The Italian newspapers made more use of the morality frame than the Dutch newspapers. However no significant difference was found between the Italian and the British newspapers. The fourth hypothesis was that the morality frame would be mostly used in the Italian media. This expectation is supported.
Concerning the powerlessness frame, a significant effect was found among the use of this frame and the three countries, F(2.297) = 6.5, p < ,05). The post-hoc test revealed that there was a significant difference between the British and the Italian articles (Mdifference= 0.18, p < .05) in the use of the powerlessness frame. The powerlessness frame was
significantly more present in the British articles compared with the Italian articles. No significant difference was found between the British and the Dutch articles concerning the use of the powerlessness frame. The expectation that the powerlessness frame would be mostly present in the British media, can be supported.
There was no significant difference found between the three countries in the use of the economic consequences frame, F(2,297) = 2.8, p = 0.06, and the attribution of responsibility frame, F(2.297) = 2.6, p =0.08. Therefore the expectations that the Netherlands would make the most use of the economic consequences frame and the attribution of responsibility frame may not be accepted.
29 Table 2.
Mean scores of the visibility of six frames in Dutch, British and Italian newspapers (mean score on a scale ranging from 0, frame not present to 1, frame present).
Frame Country Economic consequences Conflict Human interest Respon sibility Morality Powerless ness The Netherlands M N SD 0.67 100 0.36 0.13 B 100 0.38 0.09 A 100 0.19 0.35 100 0.32 0.04 C 100 0.13 0.24 100 0.36 Great Britain M N SD 0.66 100 0.38 0.60 B 100 0.42 0.28 A 100 0.33 0.30 100 0.29 0.06 100 0.20 0.33 D 100 0.38 Italy M N SD 0.56 100 0.40 0.30 B 100 0.39 0.12 A 100 0.19 0.40 100 0.32 0.12 C 100 0.28 0.15 D 100 0.34 Note. A: p < .001, B: p < .001, C: p < .05, D:p < .05
Differences in use of news frames between left-wing and right-wing newspapers Articles from six newspapers were analyzed. Three of these newspapers can be considered left-wing and three of those orientated towards the right of the political spectrum. A one-way analysis of variance (ANOVA) was carried out with the aim to find out whether newspapers with different political orientations made use of different frames The analysis showed that the newspapers only differed in their framing of the Euro crisis concerning the use of the attribution of responsibility frame, F(1.298) = 6.6, p < .05). Between the left-wing and the right-wing papers there appeared to be a significant difference. The left-wing
newspapers made more use of the responsibility frame in their framing of the Euro crisis than the right-wing newspapers (see table 3). The economic consequences frame F(1.298) = 3, p =
30 0.08, the conflict frame F(1.298) = 2.7, p = 0.10, the human interest frame F(1.298) = 1.1, p = 0.29, the morality frame F(1.297) = 0.63, p = 0.43 and the powerlessness frame F(1.298) = 0.03, p = 0.96, were not significantly different used by the left- and the right-wing
newspapers.
Table 3.
Mean scores of the visibility of six frames in Dutch, British and Italian media by political orientation of the newspapers (mean score on a scale ranging from 0, frame not present to 1, frame present). Frame Newspaper Economic consequences Conflict Human interest Respons ibility Morality Powerles sness Left-wing newspapers M N SD 0.67 150 0.38 0.34 150 0.42 0.18 150 0.28 0.40 A 150 0.33 0.06 150 0.19 0.24 150 0.34 Right-wing newspapers M N SD 0.59 150 0.38 0.36 150 0.41 0.15 150 0.24 0.30 A 150 0.28 0.08 150 0.23 0.24 150 0.39 Note: * A: p <.05
Differences in the use of news frames between the two periods of study The articles about the European crisis were analyzed in two periods around two
different Greek bailouts. A one-way analysis of variance (ANOVA) was carried out to find out whether the articles were framed in a different way in the two periods. This analysis showed that the newspapers did not significantly made more use of one of the six frames, namely the economic consequences frame: F(1.298) = 2.6, p = 0.1, the human interest frame:
31
F(1.298) = 0.1, p = 0.79, the conflict frame: F(1.298) = 0.9, p = 0.33, the powerlessness
frame: F(1.298) = 2.7, p = 0.1, the attribution of responsibility frame: F(1.298) = 1.6, p = 0.21 and the morality frame: F(1.297) = 0, p = 0.98.
Tone of the actor in the news towards the ability to solve the Euro crisis RQ1. asked whether the actors in the news were more positive or negative towards the
ability to solve the Euro crisis in the three countries. Firstly, for all the actors in the news articles, the tone towards the ability to solve the Euro crisis was analyzed. The tone was either neutral, negative, mixed or positive. The neutral scores were left out the analyses. From the 674 actors in the news, almost half, namely 321 actors did not give an evaluation. Figure 1. shows that the general tone of the actors in the news towards the ability to solve the Euro crisis is mostly negative. In the appendix examples for the tone of the actor are added.
Figure 1. The overall tone of the actors in the news towards the ability to solve the Euro
32 It was difficult to calculate the mean score of the tone of the actors in the news in SPSS, because some articles contained 5 actors and others just 2. Therefore new variables were created in which all the tones were placed together. The same was done for the background of the actors. The relationship between tone and country and tone and background of the actor was contained.
To answer the question to what extent the tone of the actors in the news towards the ability of the actors to solve the Euro crisis differs between the three countries, a one-way analysis of variance (ANOVA) was carried out. This analysis indicated that there is a significant difference between the countries concerning the tone of the actors in the news towards the ability to solve the Euro crisis, F(2.350) = 15.24, p < .001. A post-hoc test indicated that there was a significant difference in tone between Italy and the Netherlands (Mdifference = 0.41, p < .001) and between Italy and Great Britain (Mdifference = 0.62, p < .001). Between the Netherlands and Great Britain there was no significant difference concerning the tone. The tone towards the ability to solve the Euro crisis by actors in the news was the most positive in the Italian newspapers compared to the British and the Dutch newspapers (see table 4).
Table 4.
Mean score of the tone of the actor in the news towards the ability to solve the Euro crisis in three countries (mean score on a scale ranging from -1, negative to 1, positive).
Country
Tone The Netherlands Great Britain Italy Total
M N SD -0.14 A 118 0.89 -0.35 B 133 0.87 0.27 A,B 102 0.81 -0.10 353 0.89 Note: A: p < .05, B: p < .05
33 Besides the tone of the actor in the news per country, also an analysis was carried out to find out whether the Dutch, British and Italian newspapers mentioned or quoted different actors in their articles. Most of the actors in the news were ‘other EU-country actors’ (see table 5.).
Secondly, a chi-square test of independence was carried out in order to find out whether there was a significant difference in the background of the actors in the news between the Dutch, British and Italian newspapers. This test indicated that there is a
significant difference between the three countries concerning the background of the actors in the news, X² (16, N = 657) = 76.75, p < .001. Most striking differences between the countries concern the EU actors, mostly present in the news in the Netherlands (see table 5). Besides this, the domestic political actors seem to be not existent in the Italian articles. Also, the domestic government actors seem to be the least present in the British news articles. The experts seem to be most present in the Dutch news about the euro crisis, while the international level actors and the traders/investors seem to prevail in the Italian news. Moreover, the laypersons seem to be mostly present in the British news. This could be explained by the occurrence of the human interest frame, which gives a human face to the story (Semetko & Valkenburg, 2000). Because the N for the category ‘national institutions was only 1, this category was not taken into account.
34 Table 5.
The Background of the actors in the news in three countries.
Country
Background actors Netherlands
Great-Britain Italy Total
EU actors % 45.5% 25% 29.5% 100%
N 40 22 26 88
Domestic government % 44.7% 10.5% 44.7% 100%
N 17 4 17 38
Domestic political actors % 55.6% 44.4% 0% 100%
N 5 4 0 9
Organizations % 43.2% 36.4% 20.5% 100%
N 38 32 18 88
Experts % 45.9% 31.1% 23% 100%
N 28 19 14 61
Other EU-country actor % 30.9% 35.3% 33.7% 100%
N 77 88 84 249
International level actor % 20% 20% 60% 100%
N 4 4 12 20 Traders/investors % 20% 26.7% 53.3% 100% N 3 4 8 15 Laypersons % 13.5% 64% 22.5% 100% N 12 57 20 89 Total N 224 234 199 657
Finally, a chi-square test of independence was carried out to find the relation between the background of the actor in the news and the evaluation of this actor towards the ability to solve the Euro crisis. The chi-square test indicates that there is a significant difference between the different actors in the news concerning the tone towards the ability to solve the Euro crisis, X²(16, N = 347) = 84.33, p < ,001. From all actors that give negative evaluations towards the ability to solve the Euro crisis, organizations actors seem to prevail (see table 6). When looking at the mixed evaluations, the other EU-country actors seem to prevail. Finally, from all the actors that give positive evaluations towards the ability to solve the Euro crisis, the EU-country actors also seem to prevail. In the appendix, examples of these positive, mixed and negative evaluations are given.
Table. 6
Tone of the actors in the news (N=347) towards the ability to solve the Euro crisis per background of the actors
Background Actor
Column1
EU-actor Domestic Domestic 2 Organ- Experts Other EU- International- Traders/ Lay- Total Tone 5 g government Political actor izations country actor level actor investors persons
Negative % 6% 1.3% 2.6% 26.5% 19.9% 23.8% 2.0% 4.6% 13.2% 100% N 9 2 4 40 30 36 3 7 20 151 Mixed % 23.5% 10.3% 0.0% 8.8% 4.4% 44.1% 5.9% 0% 2.9% 100% N 16 7 0 6 3 30 4 0 2 68 Positive % 18% 7% 1% 13% 6% 45% 3% 6% 1% 100% N 23 9 1 17 8 58 4 7 1 128 Total N 48 18 5 63 41 124 11 14 23 347
Study 2.
Visibility of European versus other actors in Dutch, Italian and British news
Firstly, the aim was to find out how visible European actors are compared to other actors in Dutch, Italian and British news articles. Therefore, paired- sample t-tests were performed in order to find out whether the EU-actors were significantly more present in the news articles in comparison with the other five actors.
Firstly, the mean value of EU- actors present in the articles did not significantly differ from the domestic government actors, t(943) = -1.63, p = 1.03, 95% CI [-0.25, 0.02], and the other EU – country actors, t(943) = 0.06, p =0.95, 95% CI [-0.10, 0.11].
However, the paired sample t-test showed a significant difference between the EU- actors and the domestic political actors, t(943) = 7.46, p <.001, 95% CI [0.27, 0.46]. In the Dutch, British and Italian news articles, more actors seem to be EU- actors than domestic political actors.
Furthermore, the EU- actors, were significantly more visible in the Dutch, British and Italian news articles than both the Greek , t(943) = 3.63, p <.001, 95% CI [0.09, 0.30] and the international actors, t(943) = 4.70, p,<.001, 95% CI [0.14, 0.35].
Besides this, it was found that also the other-EU country actors were significantly more visible in the articles than the domestic political actors, t(943) = 8.45, p <.001, 95% CI [0.27, 0.45], the Greek actors , t(943) = 4.18, p <.001, 95% CI [0.10, 0.28] and the
37 The paired sample t-tests showed that no significant differences were found between the domestic political actors, the Greek actors and the international actors.
Table. 7
Mean scores of the visibility of six categories of actors in Dutch, British and Italian
newspapers per country (mean score on a scale ranging from 0, actor not present to 1, actor present). Actor Country EU Domestic government Domestic Political Other EU - country Greek International The Netherlands M N SD 0.54 315 1.41 0.43 315 1.29 0.11 315 0.66 0.50 315 1.18 0.30 315 0.96 0.25 315 0.92 Great Britain M N SD 0.40 318 1.24 0.87 318 2.22 0.08 318 0.54 0.55 318 1.15 0.46 318 1.16 0.28 318 0.81 Italy M N SD 0.67 311 1.47 0.66 311 1.28 0.32 311 0.76 0.55 311 1.00 0.27 311 0.99 0.35 311 1.04 Total M N SD 0.54* 944 1.38 0.65 944 1.67 0.17* 944 0.66 0.53 944 1.11 0.34* 944 1.04 0.29* 944 0.93 Note: *p<.001
Visibility of actors in Dutch, Italian and British news over time To find out whether the use of actors in the news, from March 10, 2008, until February 8,
38 First of all, concerning the use of EU- actors, the MANOVA revealed that a
significant difference exists between the periods, F(3,6.15) = 3.28, p <.05,η2 =18.46. A Bonferroni post-hoc tests shows that in the first period significantly less EU- actors were used than in both the second (Mdifference = -0.34, p<.05) and third period (Mdifference = -0.34,
p<.05).
Concerning the use of domestic government actors, also significant differences were found, F(3,17.53) = 6.65, p <.05,η2 =52.59. The Bonferroni test showed that in period 4, significantly more domestic government actors were used than in period 1 (Mdifference= 0.50, p<.05), period 2 (Mdifference= 0.57, p<.05) and period 3 (Mdifference = 0.55, p<.05).
When looking at the visibility of domestic political actors, a significant difference has been found, F(3,10.54) = 4.43, p <.05,η2 =1.87. In the fourth period, significantly more domestic political actors were used than in period three (Mdifference = 0.22, p<.05).
Concerning the use of the other EU-country actors, another significant difference has been found between the periods, F(3,25.13) = 21.58, p <.001,η2 =75.41. More other EU-country actors have been counted in the second period compared to the first (Mdifference = 0.09,p<.001) and the fourth period (Mdifference = 0.09,p<.001). The same goes for the difference between the third period and the first one (Mdifference = 0.10, p<.001) and the fourth period (Mdifference = 0.10, p<.001).
It appears the Greek actors’ visibility significantly differs between the periods,
F(3,30.12) = 30.87, p <.001,η2 =90.37. The Bonferroni post-hoc test shows that Greek actors are significantly more visible in the second and third period than the first and the fourth. More Greek actors were used in the second period compared to the first (Mdifference = .09, p<.001) and the fourth (Mdifference =.09, P<.001). Also more Greek actors were visible in
39 the articles from the third period compared to the first (Mdifference = .09, p<.001) and the fourth (Mdifference =.09, P<.001) period.
Finally, the MANOVA shows no significant differences in the use of international actors between the periods, F(3,2.27) = 2.65, p = .051 =6.8, η2 =6.81
Table. 8
Mean scores of the visibility of actors in Dutch, British and Italian news per period (mean score on a scale ranging from 0, actor not present to 1, actor present).
Actor Period EU Domestic government Domestic Political Other EU - country Greek International March 10- March 24, 2008 M N SD 0.31* 240 1.05 0.55* 240 1.70 0.16 240 0.55 0.20* 240 0.64 0.00* 240 0.00 0.30 240 1.05 April 25- May 9, 2010 M N SD 0.65* 239 1.39 0.49 239 1.44 0.18 239 0.73 0.84* 239 1.42 0.61* 239 1.45 0.42 239 0.92 February 13-February 27 2012 M N SD 0.66* 230 1.57 0.51* 230 1.18 0.06* 230 0.30 0.80* 230 1.17 0.70* 230 1.29 0.19 230 0.69 January 25- February 8 2014 M N SD 0.53 235 1.49 1.06* 235 2.14 0.28* 235 0.91 0.32* 235 0.92 0.08* 235 0.47 0.25 235 1.01 Note: *p<.001
40 Besides the change of the amount of actors counted between the periods, the
MANOVA also showed the main effects of country and political direction of the newspaper and the interaction effects of country, period and political direction on the use of frames.
When focusing on the differences between countries concerning the use of actors, it was found that the three countries differ in their use of domestic government actors,
F(2,16.24)=6.16, p<.05, η2 = 32.49. A post- hoc test with Bonferroni shows that, significantly more domestic government actors were counted in the British newspapers compared to the Dutch (Mdifference = 0,45, p<.05) newspapers. Besides this, significant differences between the countries were found, concerning the domestic political actors F(2,5.27)=12,49, p<.001,
η2
= 10.54. In the Italian news articles, significantly more domestic political actors were counted than in the Dutch (Mdifference= 0,21, p<.001) and the British articles
(Mdifference=0.24, p<.001).
Concerning the variable, left- versus right wing newspaper, also the domestic government actors seemed to differ between the countries, F(1,29.44)=6.65, p<.001, η2 = 29.44, just as the Greek actors, F(1,5.34)=5.47, p<.05, η2 = 5.34. The left-wing newspapers made significantly more use of domestic government actors (M = 0.83, SD = 0.08) than the right-wing newspapers (M= 0.48, SD= 0.08). Besides this, in the right-wing newspapers significantly more Greek actors were counted (M = 0.42, SD = 0.05) than in the left-wing newspapers (M = 0.27, SD= 0.05).
Furthermore, the interaction effects were tested. Firstly, the interaction effect of country and period on the use of the different actors, was only significant for the domestic government actors, F(6,6.32) = 2.40, p <.05,η2 =37.94. In the Netherlands, more domestic government actors were used in the fourth period (M = 1.00, SD = 0,18) compared to the first (M=0.10, SD=0.18), the second (M=0.35, SD=0.18) and the third period (M = 0.26, SD =
