Eurosceptophilia : eurosceptic and their Euro-supportive news coverage in prominent Dutch news media

Eurosceptophilia

Eurosceptic and their Euro-supportive news coverage in prominent Dutch news media

Student: Robert Danen (10420096) robert.danen@student.uva.nl

Master’s thesis

Graduate School of Communication

Master’s programme Communication Science Universiteit van Amsterdam

Supervisor: dr. L. Bos Friday January 30, 2015

ABSTRACT

In a time of steady decrease in the level of support for European integration and swift changes in the media landscape, do prominent news media still cover their news unbiased and neutral? In contrast to many claims made by scholars, the results of this content analysis of 203 news items from two prominent Dutch newspapers and one prominent Dutch online news source suggest that Dutch prominent news media do not differ in terms of their Eurosceptic and their Euro-supportive coverage. Additionally, it is suggested that there is neither a difference between news coverage of old and new news media, nor between news coverage of media with opposing political preferences when it comes to either Eurosceptic or Europhile news coverage. Consequently, no differences between identity based European news coverage and economic based European news coverage can be proven.

Keywords: old media, new media, content analysis, Euroscepticism, Europhilia, the Netherlands.

INTRODUCTION

In January 2013, the High Level Group on Media Freedom and Pluralism wrote a report on press control in which they advocated that journalists need to be restricted in terms of their European Union (EU) reporting (Van der Laan, 2013). The commission wants every EU Member State to have a council which is granted the power to impose fines to and remove journalistic status of any journalist who does not comply with their standards. In February of the same year, the European Parliament spent just over two million Euros on press monitoring in order to polish their image (Waterfield, 2013). Clearly there seems to be a certain trend that the European Commission is not fond of. This trend is widely known as Euroscepticism and

many scholars have attempted to look into the emergence and growth of this phenomenon (e.g. McLaren, 2007; Weßels, 2007). In fact, Euroscepticism is such a strong movement that actual Eurosceptic parties are being formed such as the Europe of Freedom and Democracy (EFD), the United Kingdom Independence Party (UKIP), and Order and Justice (McElroy & Benoit, 2012).

Previous studies show that in the last decade there has been a decrease in the level of support for European integration and an increase in Euroscepticism (e.g. Boomgaarden, Schuck, Elenbaas & de Vreese, 2011; Lubbers & Jaspers, 2011; Lubbers & Scheepers, 2010). Usually Euroscepticism is identified as, as McLaren (2007) describes it, a general malaise over the functioning of (European) government. To truly understand the origin of Euroscepticism it is important to know what Euroscepticism entails. Does it just consist of negative feelings toward the EU? If so, where do these feelings come from? Several scholars have divided the origins of Eurosceptic feelings into tangible, and moreover measurable, aspects (e.g. Anderson, 1998; Carey, 2002; Hooghe & Marks, 2004). One of these aspects can be defined as the identity issue, as researched by, among others, Carey (2002). It is claimed that citizens who feel their country or national sentiments are being threatened in any way project their discontent towards the EU (Anderson, 1998; McLaren, 2007). On the other hand, however, Hooghe and Marks (2004) argue that citizens who feel strongly connected to their national identity are more likely to identify with the EU. Another important element which contributes to public opinion about the EU is the economic state of a country (e.g. Anderson & Reichert, 1995; Eichenberg & Dalton, 1993; Hooghe & Marks, 2004).

The main complication with present research on Euroscepticism, though, is that Eurosceptic attitudes are so far only described and interpreted but not yet examined in news media (Boomgaarden et al., 2011). If the features of Euroscepticism are exceedingly present in news media, journalists might be accountable for the trend that is Euroscepticism

(Boomgaarden et al., 2011). Furthermore, the majority of research on Eurosceptic attitudes has been conducted during times of election. Therefore, this research will focus on Eurosceptic as well as Euro-supportive coverage in Dutch news media past election times. The choice to investigate Dutch media is based on theoretical purposes. First of all, the Netherlands is one of six countries that founded the (preliminary) EU, which was called the European Coal and Steel Community (ECSC). Secondly, they were only one of two countries that voted against a European constitution. These facts combined have caused them to be subjected to a multitude of research on Euroscepsis. There is a multitude of research on the effect of political cues on Euroscepticism (e.g. De Vries & Edwards, 2009), media attention on the EU (e.g. Lubbers & Scheepers, 2010), and news coverage of European election campaigns (e.g. De Vreese, Banducci, Semetko & Boomgaarden, 2006). However, as Boomgaarden et al (2011) mentioned, research to news media and how their content differs has not yet been conveyed. In this study, both old (newspapers) and new (online) news media will be researched. A difference between newspapers and online news sources is expected since internet news has a faster circulation with less intervention from journalists (Williams & Nicholas, 1999). Furthermore, a difference between newspapers themselves is expected because of the Netherlands’ history of pillarization (Takens, Ruigrok, Van Hoof & Scholten, 2010).

With all this in mind, the following research question can be formulated: “To what extent do prominent Dutch news media differ in terms of their Eurosceptic and their Euro-supportive coverage?”.

In order to find an answer to this research question, a content analysis of news items (N = 203) from three prominent Dutch media will be conducted.

THEORETICAL FRAMEWORK

In order to construct a solid operationalization of the concepts that are mentioned in the main research question, they should be described and explained first. This chapter will do so.

Euroscepticism

Shortly after the Second World War, six countries, including the Netherlands, founded the European Coal and Steel Community. This turned out to be the first step towards the formation of the European Union that we know today (Haas, 1968). The Maastricht Treaty of 1992 is considered the turning point of the EU (Lubbers & Jaspers, 2011; Weßels, 2007) as it established the foundation for The Economic and Monetary Union, which consequently led to the introduction the Euro. All parties involved in the signing of the treaty agreed that the future political and economical collaboration between the EU member states was the start of a journey towards a flourishing entity (Steenbergen, Edwards & De Vries, 2007). Even though the treaty and its signing was designed, implemented, and performed by elites, it was still in need of support from citizens in order for progress to take place (Anderson & Reichert, 1995). Yet, the treaty and agreements contained therein did not get the sought after backing from European citizens (Hooghe & Marks, 2009). In fact, it was the deathblow of, as Hooghe and Marks (2009) define it, “the permissive consensus that appeared to be present among EU citizens”. This culminated in 2005, when the majority of citizens in France and the Netherlands voted against a European Constitution. This made it evident that a large proportion of the population of EU member states were reluctant towards further integration of Europe as a whole (Lubbers & Scheepers, 2010). Consequently, the referendum was either postponed or cancelled in other European countries. The Dutch no was a notable disclosure, as the Netherlands had always been supportive towards European integration in the past, in particular because of their international receptiveness (Anderson & Reichert, 1995;

Eichenberg & Dalton, 1993; Lubbers & Scheepers, 2010). Overtime, compared to other EU member states, the Netherlands has since shown the largest increase in Euroscepticism (Lubbers & Jaspers, 2011). Sounds of discomfort with the appropriated power of the EU are, however, also increasingly heard from other European member states (Lubbers & Scheepers, 2005).

In recent years, a great number of studies have attempted to determine which factors may contribute to the rise of Euroscepticism, particularly in the Netherlands. Lubbers and Jaspers (2011), for example, opted that the Netherlands might have voted no against a European Constitution because of a change of generation. The generation that was supportive towards the EU, they claim, had plainly been replaced by a more Eurosceptic, younger, generation (Lubbers & Jaspers, 2011). Lubbers and Jaspers (2011) go on to claim that the topic of European integration simply was not a salient political subject in Dutch media. The salience of EU related issues, however, soon gained importance because political actors grasped the knowledge that without public consent, other referenda concerning European integration will also fail (Boomgaarden et al., 2011). Basically, Eurosceptic feelings among citizens could easily lead to stagnation or even implosion of European integration (Boomgaarden et al., 2011). According to Lubbers and Scheepers (2010), after Dutch citizens declined a European Constitution, many parties increased the volume of pro-EU messages in the media which in turn led to a decline in the level of Euroscepticism. In other words, when parties expressed a pro-EU sound, citizens did so as well (Lubbers & Scheepers, 2010).

Many other scholars claim the rise (or decline) of Euroscepticism works in much more complicated manners. They claim that the level of Euroscepticism varies across EU member states and their citizens at all times, which is why there have been various systematic attempts to examine these differences (Anderson & Reichert, 1995). Some scholars have identified two different types of Euroscepticism as a whole, namely political Euroscepticism, which focuses

on the process of resistance against the certain reduction of sovereignty of a citizen’s home country, and instrumental Euroscepticism, which puts its focus on a more practical approach (does my country benefit from being an EU member state?). However, because of the lack of research into the former kind of Euroscepticism (Lubbers & Scheepers, 2010) these will not be discussed in detail in this paper. This paper will instead focus on the, commonly agreed upon by many scholars, two dominant theoretical perspectives, which either lead to an increase or decrease in Euroscepticism among citizens: identity and economic state (Boomgaarden et al., 2011; Carey, 2002; De Vries & Edwards, 2009; Lubbers & Jaspers, 2011; McLaren, 2007; Szczerbiak & Taggart, 2003). These perspectives are often referred to as either soft factors (identity) or hard factors (economic state) and are considered the two theoretical concepts that have the most essential contribution to the clarification of Euroscepticism (Hooghe & Marks, 2005; Szczerbiak & Taggart, 2003). These two perspectives are linked to the previously mentioned political and instrumental Euroscepticism, however, they are much broader notions and cover a broader spectrum of elements. Soft factors concern threats to a EU citizen’s national culture (McLaren, 2007) whereas hard factors can be identified as the threats to the national economy of a European citizen. The current paradigm in research on Euroscepticism is that right-wing parties are more prone to make a call to arms by emphasizing the importance of national sovereignty, thus anticipating on soft factors, whereas their left-wing counterparts attempt to mobilize citizens by revealing issues that might threaten hard factors (De Vries & Edwards, 2009).

Media content and the EU

As mentioned, many scholars have conducted research on media content related to the EU. For instance, whether or not, and if so in which way, news about European politics is framed has received plenty attention (e.g. De Vreese, 2007; Semetko & Valkenburg, 2000). Other

popular research topics concern whether or not there is a difference between different types of media (television, radio etc.) (e.g. Bucy, 2003; Maier, 2010), the way in which the media cover European politics (e.g. Gleissner & De Vreese, 2005; Machill, Beiler & Fischer, 2006), and the influence of media use on public opinion concerning the EU (e.g. Eichenberg & Dalton, 1993; Hooghe & Marks, 2004; McLaren, 2002). To top it off, many of these studies have been partly replicated and put into context by means of comparative studies (e.g. De Vreese, 2001; Taggart, 1998). Interestingly, the content of media has been claimed to have numerous effects on citizens, for instance on their voting behaviour (DellaVigna & Kaplan, 2007). Therefore, it would be interesting to take into account what effects media could have on Euroscepticism.

Effect of media on Euroscepticism

If Dutch media outlets still are somewhat pillarized, as mentioned in the introduction, and thus either leaning towards Euroscepticism or Europhilia, this could have an effect on citizens and their attitudes towards the EU. It is argued that media coverage of EU-related issues sets the public agenda on these issues (e.g. De Vreese, 2007). Furthermore, it seems as though the media do not only influence what issues citizens think of, but also how they think about, these issues (De Vreese, 2007). Additionally, media especially appear to influence the level of political cynicism, which in turn is related to an increase in Euroscepticism (Bennett, Rhine, Flickinger & Bennett, 1999; De Vreese, 2007; Krouwel & Abts, 2007; Van Spanje & De Vreese, 2014).

Soft factors

Research on the role of threats to EU citizen’s identity, or nationalist sentiments as Lubbers (2008) calls it, has only taken shape in more recent years (Lubbers, 2008). McLaren (2002)

claims that many scholars adumbrate EU commonalty as a competition between various social classes which summons existing national sentiments. This in turn causes citizens who are attached to their culture, nationality, or identity to discard any kind of integration of Europe as a threat to their identity and such messages have been shown to have a significant negative effect on EU support (Carey, 2002; Hooghe & Marks, 2005). This includes anti-immigrant attitudes. In fact, anti-anti-immigrant attitudes have been shown to play a role in several (experimental) referenda votes (De Vreese & Boomgaarden, 2005). This is odd. The EU was originally founded to unite European countries after the Second World War and to aid in European trade. This calls for a union characterized by open borders, which implies no EU citizens could be considered immigrants within the EU. Furthermore, one would not expect that a union as apparently open-minded (in terms of ancestry) as the EU would be anxious when it comes to people from a different nationality, albeit non-European. Because anti-immigrant attitudes appear to be such a significant element of Euroscepticism based on national sentiments, right-wing politicians and parties often attempt to intensify Euroscepticism among citizens (Hooghe & Marks, 2004). Of course pro-EU politicians on the other end of the political spectrum use the EU in their favour as well. Hooghe and Marks (2004) argue that in countries with a Europhile elite, national identity is positively related to support for European integration. Another element that can strengthen one’s national sentiments and therefore increase Euroscepticism among European citizens is the redistribution of (political) authority. When countries pass on authority from their national level to the European level it is often interpreted as an aforementioned threat to one’s identity (McLaren, 2007). Again, when considering the reasons for the foundation of the EU, this seems peculiar. Within the Netherlands the size of the country also matters. Some Dutch citizens fear that whenever the EU expands, a small country such as their own will be completely overruled (Lubbers, 2008). Taking all this into consideration, it is expected that

the more Dutch media outlets cover stories that either talk about harming or supporting national identity, such as anti-immigration issues, or shifts in political power, the level of Euroscepticism among Dutch people will alternate. Consequently, it is of importance to study media content. The following research question can be formulated accordingly:

RQ1: To what extent are soft factors present in prominent Dutch news media?

Hard factors

The EU started off as a trade union, created to strengthen the economies of the founding member states as well as easing processes involved with international trade (Anderson & Reichert, 1995). Of course, overtime, the economic advantages (and disadvantages) started playing a bigger part as the EU expanded. The economic state of individuals is not the only way the economy affects EU citizens, the economic conditions of EU member states also reflect upon the EU itself (Eichenberg & Dalton, 1993). This causes citizens to not just be receptive to plights that have an effect on them personally, but also to situations that affect their collective economy (Anderson, 1998). Consequently, citizens who have trust in the future when it comes to the economy (both on an individual level as well as on a collective level) are more prone to be supportive of the EU, whereas citizens who do not have this trust are prone to Euroscepticism (Hooghe & Marks, 2004). Logically, countries that are on the receiving end of EU spending have more Europhile citizens whereas citizens of countries that can be considered so-called donor countries will have a more Eurosceptic mindset (Anderson & Reichert, 1995; Eichenberg & Dalton, 1993; Hooghe & Marks, 2004; Hooghe & Marks, 2005; Lubbers & Jaspers, 2010). The Netherlands can be considered a donor country. In fact, the Netherlands has been the main financial contributor in the EU for several consecutive years (CBS, 2010). For example, in 2008 the remittance per Dutch citizen was 268 Euros, the

highest number of all member states (CBS, 2010). This fact can make one doubt the conclusions of Lubbers (2008) which state Dutch citizens fear their country could be overruled by the EU, as mentioned in the previous paragraph. Clearly, the Dutch have a prominent position within the EU landscape. Additionally, the Netherlands received the least amount of money of all EU member states (CBS, 2010). Consequently, it is expected that the majority of EU coverage in Dutch media will be about the economic state of the Netherlands as opposed to soft factors.This assumption leads to the following research question:

RQ2: To what extent are hard factors present in prominent Dutch news media?

Differences between media

In the digital era we now live in scholars often divide media into two groups; old media (newspapers, television) and new media (internet) (Maier, 2010). The amount of people that get their news from online news sources has vastly expanded over the last few years and old media appear to be a dying breed. Maier (2010) even claims that since 2008, online news gathering has overtaken all other modes of news gathering, except for television. This trend challenges journalists of old media to be on a continuous quest to find ways to bind people to their news medium. This is attempted in a variety of manners, such as adding certain frames to news items or by giving an article a sensational angle (Hendriks Vettehen, Nuijten & Peeters, 2008). Hence arises, as expected in a commercial market, competition between various old media. Additionally, the growth of new media, caused by the migration of users from old media to new media, contributes to the economic turmoil existing among old media outlets. Consequently, expenses are cut down and journalist are being let go. This in turn causes a shift among old media outlets in which they resort to copy/pasting news items

directly from news agencies such as AP and Reuters (Maier, 2010; Manning, 2001). So on one hand old media outlets have to differentiate themselves from other media in order to both gain and retain readers, but there are less journalists to do so which might cause them to replicate items from news organisations. Although these circumstances appear to be quite paradoxical, there is no evidence that this current shift actually affects the news coverage of both media types (Maier, 2010). If new media would report in the exact same manner as old media this would mean nothing more than an overall transformation from paper to digital. That implies that by keeping the manner in which old media report as their own, newspapers could well retain their readers as well as gain new ones. However, this would require a separate identity for each media outlet belonging to old media, which can be achieved by manually influencing or complementing news stories (Manning, 2001).

Left-wing and right-wing media

Among scholars it is common knowledge that Western-European party systems have evolved around a dominant left-right dimension in which socialist parties are in favour of economic state intervention and more conservative parties are in favour of free market politics (e.g. Bartolini & Mair, 1990). For a long time, scholars dissociated European conflicts from this dimension, but they now seem to agree that Euroscepticism does, indeed, take place within this dimension (De Vries & Edwards, 2009). Euroscepticism appears to be more present on the right side of the political spectrum whereas their left-wing counterparts are more Europhile (De Vries & Edwards, 2009; Hooghe & Marks, 2004; Taggart, 1998). The media system in the Netherlands was established during a period of pillarization, a time in which society was divided by both religious as well as political interest groups (Takens et al., 2010). Even though depillarization has obliterated most of this division, some media still have the tendency to lean towards a certain side of the political spectrum (Semetko & Valkenburg,

2000; Takens et al., 2010). The aforementioned competition in the present news market has increased these tendencies, thus it is expected that some media outlets contain more Eurosceptic news coverage than others (Takens et al., 2010). Consequently, the following hypotheses and research question can be formulated:

H1: Old media will contain more Eurosceptic news coverage than new media.

H2: Right-wing media will contain more Eurosceptic news coverage than left-wing media. RQ3: To what extent do left-wing media and right-wing media differ in their use of soft and hard factors?

Oddly, an obtrusive trend which can be observed in present day research is the obvious focus on the negative aspects of European integration. No scholar has attempted to expose the presence of Europhile news coverage, for example. However, negativity, in this case Euroscepticism, can only be put into perspective when it is compared to positivity, in this case Europhilia. Hence, this research will also focus on positive attitudes and aspects of European integration. Therefore, the following research question can be formulated:

RQ4: To what extent is Europhile news coverage present in old and new media and left-wing and right-wing media?

METHOD

The aforementioned research question of this paper will be answered by means of a systematic quantitative content analysis. With a content analysis one can systematically structure, and describe the meanings of, large amounts of text with focus on latent contents

and context (Burla et al., 2008). As this study will focus on text from newspapers as well as text from news websites, a content analysis is the most appropriate method of research.

Sample

News websites can be classified as platforms where a selection of editorial content is offered, usually copied from press agencies or linked from an external source, often oriented towards a large audience (Deuze, 2003). For this research, Dutch independent news website nu.nl has been selected. The choice for nu.nl is an obvious one. First of all, nu.nl is the largest and most popular news website in the Netherlands in terms of its unique visitors. Secondly, news websites affiliated with newspapers generally place the very same content as their paper counterparts, which is why no online newspaper has been selected (Sanoma, 2013; Tewksbury & Rittenberg, 2012). The newspapers selected for this study are De Telegraaf and Volkskrant, two newspapers that are in the top three of most read newspapers in the Netherlands (HOI, 2015). Not only are they among the top of most read newspapers, they also tend to lean to the right and the left side of the political spectrum respectively (Semetko & Valkenburg, 2000). This aids the current study’s purpose because, as mentioned in the theoretical framework, it is generally claimed that in liberal market systems, such as the one in the Netherlands, media leaning towards the right side of the political spectrum often oppose European integration and their left-wing counterparts do the opposite (Hooghe & Marks, 2004). By selecting three large and popular sources of media results may be easier to generalize at a later stage, which benefits this study’s validity.

The sample of this study consist of news articles from the aforementioned media published between September 1st 2014 and December 1st 2014. This time period is chosen for several reasons. Firstly, there have been no European Elections during the selected three months, which is in line with this study’s purpose. Secondly, the focus of this research lies on

present day news coverage and not on news coverage from the past which is why these recent three months have been selected.

Because this study focuses on news articles in which the EU is the main subject, the selected sample of news articles consisted of all EU related articles of all three media published in the predetermined period of time. To obtain only EU related articles, keywords were used to aid the search. The following three keywords have been used in this particular process: Europese Unie (European Union), EU and Europese Commissie (European Commission). European Commission was added to the keywords because of the high probability that articles mentioning the European Commission are EU related. This lead to a total sample of 844 newspaper articles, which can be subdivided into De Telegraaf (N = 408) and Volkskrant (N = 436) and accompanied by news items from nu.nl (N = 83). Now it may seem odd that nu.nl has far fewer results than both newspapers combined. However, nu.nl adds search tags to every news article they publish. This both facilitates visitors of nu.nl when they want to look up news about a certain subject to find the desired articles as well as nu.nl’s search engine optimization (the visibility of a website in search engines such as Google). This lead to all nu.nl articles being specifically related to news about the EU, whereas articles from both newspapers also contained articles which mentioned the EU merely in a trivial or superficial manner (less specific). In order to equalize the amount of newspaper articles and online articles, systematic sampling was carried out which ultimately led to a sample of De Telegraaf (N = 81) and Volkskrant (N = 87) and, as mentioned, nu.nl (N = 83). For a more detailed overview of the sampling procedure, especially with regard to how the systematic sampling has been executed, please consult appendix A.

The data collection for this study went through Lexis Nexis, an online news article database often consulted by scholars. For data collection of online news articles, nu.nl’s own archive was used. For this content analysis, only news items concerning the EU have eventually been included. Finally, after removing duplicates from the nu.nl sample, the total amount of EU related articles was as follows: Telegraaf (N = 65), Volkskrant (N = 60) and nu.nl (N = 78). Table 1.1 shows a representation of the sample data for each medium. Again, please consult appendix A for a more detailed overview of the sampling process.

<INSERT TABLE 1.1 HERE>

To select only those articles from the sample of newspaper articles that were actually EU related, a filter variable was added at the very beginning of the codebook in order to filter out those articles that did not comply. After several identification variables were coded and a variable was considered EU related, the next step in the coding process was to code various control variables which measured the prominence of the article. Aspects that were considered are, for instance, the length of the article, whether or not the article was a regular news item or an opinion piece, and the topic of the article. Furthermore, the overall tone of the article had to be coded. This variable measured whether the tone of the article was either positive, neutral, or negative towards EU related parties or actors present in the article.

Soft factors

The dependent variable in this study is Euroscepticism. As acknowledged in the theoretical framework, Euroscepticism consists of two dominant theoretical perspectives: soft and hard factors. In order to find out whether or not soft factors were present in the articles a total of 25 dichotomous variables have been drafted which could either be coded with value ‘0’ (no) or

‘1’ (yes). This may seem an abundance of variables but they only cover a total of eight aspects which are considered the main elements of the theoretical concept of soft factors (De Vreese & Boomgaarden, 2005; Hooghe & Marks, 2004; Lubbers, 2008; McLaren, 2007). The answering scale has been kept dichotomous for analysis purposes. For example, one element of soft factors is redistribution of (political) authority. This element has been measured with three variables. One variable considered whether or not the article mentioned any kind of redistribution of political authority, whereas the other two variables considered whether or not this element was treated in a positive or in a negative manner. The two latter variables only had to be coded if the first variable was coded with a ‘1’ (yes). The other elements of soft factors are as follows: indications that Europeans share a common tradition, culture and history, mentioning of people who are proud to be a EU citizen, the importance of national sovereignty, the enrichment of EU member states caused by European integration, immigration, the rate of immigration, and the position of the Netherlands within the EU’s political sphere. In the process of creating these variables, variables from two existing studies have been adopted and fitted, simply because they have been proven to be excessively suitable for measuring what are considered to be aspects of soft factors (Boomgaarden et al., 2011; Schuck, Xezonakis, Elenbaas, Banducci & De Vreese, 2011). All 25 variables can be consulted in the codebook, which can be found in appendix B.

Hard factors

In order to find out whether or not hard factors were present in the articles a total of 11 dichotomous variables have been drafted which either could be coded with value ‘0’ (no) or ‘1’ (yes). Similar to coding soft factor elements, not all variables measured a different element of the hard factors. In fact, the 11 created variables measured three elements of hard factors; the presence, evaluation and responsibility of economic conditions, whether or not these

conditions are enriched or threatened by the EU, and whether or not there is any mention of the Netherlands as being either on the receiving or on the donating end of EU spending. For the creation of these variables, again variables from an existing study have been adopted and fitted because they have been proven to be suitable for measuring what are considered to be aspects of hard factors (Schuck et al., 2011). All 11 variables concerning hard factors can be consulted in the codebook, which can be found in appendix B.

Intercoder reliability

All coding was done by one coder. However, to make sure that any other person would code the data in accordance to what the codebook is supposed to measure (possibly in future research), and to aid the validity of this present study, an external person has been consulted in order to attain intercoder reliability. 45 articles have been selected (N = 15 for each of the three sources) and, after thorough training, have been coded separately by the two coders. To calculate whether or not the two coders were in compliance with the codebook, a reliability analysis has been performed, which is expressed in Krippendorff’s alpha (KALPHA) values. Generally, KALPHA values have a lower limit of .60, which is considered the minimal accepted reliability (also known as substantial reliability). However, the aim is to use variables with a KALPHA value of .80 or higher, which is considered ‘near perfect’ reliability. When a KALPHA value is 1 there is perfect reliability, in which case the coders were in agreement on all variables. After the 45 articles were coded the first time, no KALPHA values lower than .60 were observed. However, a total of five variables had KALPHA values between .60 and .80. Lastly, an astonishing 24 variables were all coded with the same value, which, in this instance, meant that they did not appear to be present in any news article. When this is the case, KALPHA values appear as undefined due to invariant variables. However, when this is the case it does not mean these variables cannot be used. As

shown in appendix C, the percent agreement on each of the variables is over 93 per cent at all times, which is a clear sign of coder agreement. In conclusion, the overall reliability was satisfactory. All KALPHA values can be consulted in appendix C.

Analysis

This paragraph will cover the analyses which have been performed in order to formulate a well founded answer to the research question, namely Mokken Scale analysis, independent samples T-test, and regression analysis.

Mokken Scale analysis

When handling binary (dichotomous) variables, it is not common to perform the generally used factor analysis. Instead, Mokken Scale analysis has been used in order to find out whether or not the variables in the codebook actually measure soft and hard factors respectively. By use of the Coefficient of Homogeneity (H) of Loevinger, values are calculated which represent the strength of a certain scale (Mokken, 1971). Mokken Scale analysis has a lower limit of .30, in which case at least a weak scale is present. Because this study focuses on both positive and negative aspects of EU attitudes, it has been attempted to create a total of four scales. They have been divided as follows: one scale containing positive aspects on soft factors, one scale containing negative aspects on soft factors, one scale containing positive aspects on hard factors, and one scale containing negative aspects on hard factors.

Soft factors

Using the item selection procedure with a lower limit of H = .30, the Mokken Scale analysis has shown that no scale can be created with any of the variables used to measure soft factors (see table 1.2). This can partly be explained by the fact that soft factors did not appear to be of

frequent occurrence in the news item coded for this study. Another, more notable, explanation of the fact that no scale could be created for soft factors is that when an element of soft factors was present in a news item, it was never in combination with another element of soft factors. However, for analytical purposes and to attempt to answer the created research questions and hypotheses formulated in this study, two scales were created anyway. One consisting of all positive elements of soft factors, and one consisting of all negative elements of soft factors.

Hard factors

Using the item selection procedure with a lower limit of H = .30, the Mokken Scale analysis has shown that a scale can be created with all six variables concerning the positive aspects of hard factors. This scale is considered strong (H = .64).

Using the item selection procedure with a lower limit of H = .30, the Mokken Scale analysis has also shown that a scale can be created with five of six variables concerning the negative aspects of hard factors. This scale is considered strong (H = .56). Please consult table 1.2 for all H-scores.

<INSERT TABLE 1.2 HERE>

Independent-samples T Test

By means of an independent-samples T Test, one can compare means of two groups on a continuous dependent variable (Elvers, n.d.). In this study, two T tests have been carried out. One in order to explore whether or not the two media types (old and new) differ in their use of soft or hard factors and Eurosceptic news coverage as a whole, another to explore whether or not political tendency plays a role in the amount of soft or hard factors that are mentioned in news items or in the amount of Eurosceptic news coverage as a whole. To perform tests on

Euroscepticism as a whole a new variable has been computed, Eurosceptic, which consist of both scales measuring the negative aspects of soft and hard factors. Additionally, to test RQ4: To what extent is Europhile news coverage present in old and new media and left-wing and right-wing media? a variable ‘Europhile’ has been created by adding the two scales that measure the positive aspects of soft and hard factors. The descriptives of these variables can be found in table 1.3.

<INSERT TABLE 1.3 HERE>

Regression analysis

A regression analysis is performed in order to predict the scores of the dependent variables (soft and hard factors) on the value of the independent variables (the type of media and political tendency). This way, an attempt is being made to predict the amount of Euroscepticism in news items based on the type of media a news item is publish in and based on the political favour of the medium.

Control variables

In order to verify data for attributes which could be customary explanations for significant results, several control variables will be added to the regular regression analysis. For instance, based on common sense it can be expected that the length of a news item can influence the presence of soft or hard factors: the more words an article contains, the more possibilities for such aspects to be present. Furthermore, it can be expected that the type of news item influences the presence of soft or hard factors: based on common sense one can expect opinion pieces to contain more of such aspects. Moreover, the overall tone of the news item (negative (-1), neutral (0) or positive (1)) is expected to have an effect on the amount of soft

and hard factors present in the news item and the prominence of the news item (whether or not an article was presented on the front page of a newspaper) is too: articles presented on the front page are supposed to draw the reader’s attention (Schuck et al., 2011; Strömback & Dimitrova, 2006) and it is expected that use of soft and hard factors does so. Lastly, the presence of Dutch, EU or non-EU level political actors will be considered as it is expected that these aspects influence the amount of soft and hard factors used in a news item (Schuck et al., 2011). Please consult table 1.4 for an overview of the control variables and their descriptives.

<INSERT TABLE 1.4 HERE>

RESULTS

In this paragraph, the analysis which are described in the previous paragraph, namely independent-samples T Test and regression analysis, will be conducted in order to answer the four constructed hypotheses.

Descriptive statistics

As shown in the aforementioned table 1.3, RQ1 “To what extent are soft factors present in prominent Dutch news media?” and RQ2 “To what extent are hard factors present in prominent Dutch news media?” can be answered with a simple yet clear ‘no’. It is obvious that both soft factors containing negative attitudes (M = 0.02, SD = 0.04) and positive attitudes (M = 0.02, SD = 0.05) as well as hard factors containing negatives attitudes (M = 0.05, SD = 0.14) and positive attitudes (M = 0.04, SD = 0.12) are hardly present in prominent Dutch news media.

When performing an independent-samples T Test, it is necessary to verify whether or not the four created scales are distributed normally (De Vocht, 2013). While consulting a frequency table, it became clear that this was not the case for the variable SoftPOS (M = 0.02, SD = 0.05, skewness = 4.26, kurtosis = 22.63), nor was it the case for the three other scaled variables, namely: SoftNEG (M = 0.02, SD = 0.04, skewness = 2.67, kurtosis = 6.61), HardPOS (M = 0.04, SD = 0.12, skewness = 4.26, kurtosis = 23.84), and HardNEG (M = 0.05, SD = 0.14, skewness = 3.46, kurtosis = 14.48). Therefore, a log transformation has been conducted in order to attempt to improve the distribution of all four variables. This caused the four variables to improve as follows: SoftPOS (M = 0.02, SD = 0.05, skewness = 3.88, kurtosis = 18.43), SoftNEG (M = 0.01, SD = 0.04, skewness = 2.56, kurtosis = 6.05), HardPOS (M = 0.04, SD = 0.1, skewness = 3.44, kurtosis = 14.57), and HardNEG (M = 0.04, SD = 0.11, skewness = 2.94, kurtosis = 9.22). Although the skewness is still not conducive, analysis with said variables was still carried out. This decision is based on the commonly agreed upon consideration that not conducting analysis due to a substantial skewness is an arbitrary rule.

The first independent-samples T Test, which is carried out in order to formulate an answer to hypothesis H1 “Old media will contain more Eurosceptic news coverage than new media”, shows, with 95% certainty, that old media (M = 0.01, SD = 0.04) and new media (M = 0.01, SD = 0.04) do not significantly differ in their use of Eurosceptic soft factors in their news items, t(201) = -0.49, p = .625, 95% CI [-0.01, 0.01]. The first independent-samples T Test also shows, with 95% certainty, that old media (M = 0.04, SD = 0.11) and new media (M = 0.04, SD = 0.11) do not significantly differ in their use of Eurosceptic hard factors in their news items, t(201) = 0.06, p = .953, 95% CI [-0.03, 0.03]. Furthermore, the first independent-samples T Test shows, with 95% certainty, that old media (M = 0.03, SD = 0.06) and new

media (M = 0.03, SD = 0.06) do not significantly differ in their use of Eurosceptic news coverage overall, t(201) = -0.10, p = .919, 95% CI [-0.02, 0.02].

The second independent-sample T Test was performed to test hypothesis H2 “Right-wing media will contain more Eurosceptic news coverage than left-“Right-wing media” and to answer RQ3 “To what extent do left-wing media and right-wing media differ in their use of soft and hard factors?”. This independent-samples T Test shows, with 95% certainty, that left-wing media (M = 0.01, SD = 0.04) and right-wing media (M = 0.01, SD = 0.04) do not significantly differ in their use of Eurosceptic soft factors in their news items, t(123) = 0.19, p = .854, 95% CI [-0.01, 0.01]. Furthermore, it can be concluded that, with 95% certainty, left-wing media (M = 0.02, SD = 0.07) and right-left-wing media (M = 0.06, SD = 0.14) do not significantly differ in their use of Eurosceptic hard factors in their news items, t(123) = -1.82, p = .072, 95% CI [-0.08, 0.00]. This independent-samples T Test also shows, with 95% certainty, that left-wing media (M = 0.02, SD = 0.04) and right-wing media (M = 0.04, SD = 0.07) do not significantly differ in their use of Eurosceptic news coverage overall, t(123) = -1.67, p = .097, 95% CI [-0.04, 0.00].

Regression analysis

Euroscepticism

The regression model (see model 1a and model 1b in table 2.1.) with the amount of Euroscepticism in news items as dependent variable and the type of media (old or new), the weight variable for average length of the news item, the prominence of the news item, the type of news item, the presence of Dutch actors, the presence of EU actors, the presence of foreign actors and the overall tone of the news item as independent variables shows a significant result F(8, 194) = 3.52, p = .001. This means that the regression model can be used to predict the amount of Euroscepticism in news items, but the strength of the prediction can

be considered very weak: a mere thirteen per cent of the variation in the amount of Euroscepticism in a news item can be predicted based on the type of media (old or new) the news item is in (R² = .13).

The presence of a European (or EU level) political actor b* = -0.18, t = -2.47, p = .015, 95% CI [-0.05, -0.01] and the overall tone of a news item b* = -0.21, t = -2.51, p = .013, 95% CI [-0.08, -0.01] both have a significantly weak association with the amount of Euroscepticism in a news item. For these effects, it is assumed that the other independent variables are held constant. No significant effects have been found for the type of media.

The increase of the predicted amount of Euroscepticism in a news item (B) cannot be interpreted for any of the selected independent variables. This is due to the performance of a log transformation at the beginning of the analysis. An overview of the requested regression models, in which all scores are shown, can be consulted as model 1a and model 1b in table 2.1.

The regression model (see model 2a and model 2b in table 2.1) with the amount of Euroscepticism in news items as dependent variable and the political tendency of media (right-wing and left-wing), the weight variable for average length of the news item, the prominence of the news item, the type of news item, the presence of Dutch actors, the presence of EU actors, the presence of foreign actors and the overall tone of the news item as independent variables shows a significant result F(9, 193) = 3.46, p = .001. This means that the regression model can be used to predict the amount of Euroscepticism in news items, but the strength of the prediction can be considered very weak: a mere fourteen per cent of the variation in the amount of Euroscepticism in a news item can be predicted based on the political tendency of the medium the news item is in (R² = .14).

As in the previous regression analysis, the presence of a EU political actor b* = -0.17, t = 2.33, p = .021, 95% CI [0.05, 0.00] and the overall tone of the news item b* = 0.2, t =

-2.41, p = .017, 95% CI [-0.08, -0.01] appeared to have a significantly weak association with the amount of Euroscepticism in a news item. For these effects, it is assumed that the other independent variables are held constant. No significant effects have been found for the political tendency of media. Due to the performed log transformation, the increase of the predicted amount of Euroscepticism in a news item (B) cannot be interpreted for any of the selected independent variables. An overview of the requested regression models, in which all scores are shown, can be consulted as model 2a and model 2b in table 2.1.

<INSERT TABLE 2.1 HERE>

Europhilia

To formulate a reasonable answer for RQ4 “To what extent is Europhile news coverage present in old and new media and left-wing and right-wing media?” two regression analyses are performed to see if it is possible to predict the amount of Europhilia in news items. One for the type of news media and one for the political tendency of news media.

The regression model (see 1a and model 1b in table 2.2) with the amount of Europhilia in news items as dependent variable and the type of media (old or new), the weight variable for average length of the news item, the prominence of the news item, the type of news item, the presence of Dutch actors, the presence of EU actors, the presence of foreign actors and the overall tone of the news item as independent variables shows a significant result F(8, 194) = 2.29, p = .02. This means that the regression model can be used to predict the amount of Europhilia in news items, but the strength of the prediction can be considered very weak: a mere nine per cent of the variation in the amount of Europhilia in a news item can be predicted based on the type of media (old or new) the news item is in (R² = .09).

Only the presence of a European (or EU level) political actor b* = -0.2, t = -2.56, p = .011, 95% CI [-0.09, -0.01] has a significantly weak association with the amount of Europhilia in a news item. For this effect, it is assumed that the other independent variables are held constant.

Once again, due to performing a log transformation at the early stages of the analysis process, the increase of the predicted amount of Europhilia in a news item (B) cannot be interpreted for the selected independent variable. An overview of the requested regression models, in which all scores are shown, can be consulted as model 1a and model 1b in table 2.2.

The regression model (see model 2a and model 2b in table 2.2) with the amount of Europhilia in news items as dependent variable and the political tendency of media (right-wing and left-(right-wing), the weight variable for average length of the news item, the prominence of the news item, the type of news item, the presence of Dutch actors, the presence of EU actors, the presence of foreign actors and the overall tone of the news item as independent variables shows a significant result F(9, 193) = 2.06, p = .035. This means that the regression model can be used to predict the amount of Europhilia in news items, but the strength of the prediction can be considered very weak: a mere nine per cent of the variation in the amount of Europhilia in a news item can be predicted based on the political tendency of the medium the news item is in (R² = .09).

The presence of a EU political actor b* = 0.19, t = 2.51, p = .013, 95% CI [0.09, -0.01] appears to have a significantly weak association with the amount of Europhilia in a news item. For this effect, it is assumed that the other independent variables are held constant. Due to the performed log transformation, the increase of the predicted amount of Euroscepticism in a news item (B) cannot be interpreted for any of the selected independent

variables. An overview of the requested regression models, in which all scores are shown, can be consulted as model 2a and model 2b in table 2.2.

<INSERT TABLE 2.2 HERE>

CONCLUSION AND DISCUSSION

This research has attempted to formulate an answer to the question whether or not prominent Dutch news media differ in terms of their Eurosceptic and their Euro-supportive coverage. To make Euroscepticism and Europhilia more tangible and measurable, focus of this study has been put on soft and hard factors of sentiments towards European integration. A content analysis has been conducted which included a total of 203 news articles from three prominent Dutch media, namely nu.nl, an online Dutch news website, De Telegraaf, a Dutch newspaper leaning towards the right side of the political spectrum, and, in terms of political tendency, Telegraaf’s counterpart De Volkskrant. The analyses have shown that prominent Dutch news media do not differ in terms of their Eurosceptic and their Euro-supportive coverage. A regression analysis has shown a few significant results, namely that the amount of Euroscepticism, and the amount of Europhilia for that matter, can be predicted based on the type of media and based on the political tendency of media. However, these results cannot be interpreted due to the performance of a log transformation, which in turn was performed due to an indifferent skewness of the scales. A possible explanation for the skewness of both scales being off can be sought in the difference of variables measuring a scale. Overtime, during coding, it became rather clear that whenever a variable was coded as being present in a news item, this would very few times be in accordance to other variables within the same scale. Generally speaking, the measured variables hardly occurred in the coded news items. Hence, it is recommended for future research to attempt to seek conformity within scales of

variables. Because this present study has attempted to create two entirely new scales based on existing theory, it may be necessary in future studies to conduct even more thorough research before creating scales, not in the least because one of the two created scales was purely created for analytical purposes; scale analysis has shown the used variables were not a strong scale to begin with. Another solution may be to utilize existing scales, such as the ones from Boomgaarden et al. (2011). Lastly, the two scales created for this study could perhaps be divided into several, smaller, scales.

Because of the exploratory nature of this research, a content analysis has been conducted. In this way, the presence of Euroscepticism or Europhilia in news items can well be indicated and delineated. The fact no significant results, and thus no differences between prominent Dutch news media, have been found is unforeseen. First of all, this study has shown that elements of Euroscepticism are not at all present in prominent Dutch media. However, Lubbers and Jaspers (2011) claim the Netherlands has shown the largest increase in Euroscepticism of all EU member states. When considering the study of De Vreese (2007), in which he argues that media coverage of EU-related issues set the public agenda on these issues, one can thus expect that especially Dutch media would consist of somewhat Eurosceptic news coverage. Possibly the negative attitude of Dutch citizens towards European integration does not stem from media coverage and perhaps other factors which could be responsible for the Euroscepsis among Dutch citizens should be researched. Support for this claim can be found in the study of Lubbers and Scheepers (2010), in which they argue that when parties expressed a pro-EU sound, citizens did so as well. It could well be possible that political parties and its actors are the actual setters of the agenda. Another possible explanation lies in the fact that the amount of news items considered for this content analysis (N) was rather low due to time constraints. Moreover, the time period in which the considered news items were published was outside of election times. Even though this was in accordance

with the goal of this study, it may still clarify the absence of any such results. In any case, the results from this present study are well-considered and should be duly noted.

Peculiarly, during coding, a specific negative economic event occurred. As it turned out, the Netherlands was ordered to pay a 642 million Euro budget bill to the EU (ANP, 2014). One could say this would pave the way for any (specifically right-wing) journalist to publish Eurosceptic news items, not in the least because citizens of countries that can be considered so-called donor countries will have a more Eurosceptic mindset (e.g. Anderson & Reichert, 1995; Eichenberg & Dalton, 1993). However, the results prove no such thing. As it goes to show, prominent Dutch media is not one to speak of unity and recovery after a negative occurrence. This, however, could be due to the Dutch media system as described by Hallin and Mancini (2004), since their American counterparts have been shown to propose unity after unfortunate events (Cali, 2009).

Lastly, it has been argued that the Europhile generation has been replaced by a more Eurosceptic, younger, generation (Lubbers & Jaspers, 2011). This could easily lead up to a great difference in the way old media (newspapers) report opposed to new media (internet) because, as claimed by several scholars, the audience of online news sources is younger than the audience of old media (e.g. Bucy, 2003). However, no such results have been found. Perhaps the new, younger, generation of journalists are able to objectively present news facts, as is desired among professional journalists (Shehata, 2010).

In order to construct possible significant results and to further analyze the trend that is Euroscepticism, future scholars should, for starters, include more news items in their research. Furthermore, it is recommended that a larger time period will be selected. Additionally, when significant results do suggest presence of Euroscepticism or Europhilia in news items, comparative research overtime could be conducted to spot trends. Consequently, comparative research involving various countries can map the Euroscepticism phenomenon.

In conclusion, it appears as if the European Parliament could have used their two million Euro press monitor investment in a much better way. Perhaps they can gain a vast amount of Dutch Euro fans by covering a part of that multi million Euro budget bill.

LITERATURE

Anderson, C. J. (1998). When in doubt, use proxies: Attitudes toward domestic politics and support for European integration. Comparative political studies, 31(5), 569-601.

Anderson, C. J., & Reichert, M. S. (1995). Economic benefits and support for membership in the EU: A cross-national analysis. Journal of Public Policy, 15(3), 231-250.

ANP (2014, October). Dijsselbloem ‘zeer onaangenaam verrast’ over naheffing EU. Retrieved from

http://www.nu.nl/politiek/3911398/dijsselbloem-zeer-onaangenaam-verrast-naheffing-eu.html.

Bartolini, S., & Mair, P. (1990). Policy competition, spatial distance and electoral instability. West European Politics, 13(4), 1-16.

Bennett, S. E., Rhine, S. L., Flickinger, R. S., & Bennett, L. L. (1999). "Video Malaise" revisited: Public trust in the media and government. The Harvard International Journal of Press/Politics, 4(4), 8-23.

Boomgaarden, H. G., Schuck, A. R., Elenbaas, M., & de Vreese, C. H. (2011). Mapping EU attitudes: Conceptual and empirical dimensions of Euroscepticism and EU support. European Union Politics, 12(2), 241-266.

Brinegar, A., Jolly, S., & Kitschelt, H. (2004). Varieties of capitalism and political divides over European integration. In Marks, G., & Steenbergen, M. R. (Eds.) European

Integration and Political Conflict (pp. 62-89). Cambridge: Cambridge University Press. Bucy, E. P. (2003). Media credibility reconsidered: Synergy effects between on-air and online

news. Journalism & Mass Communication Quarterly, 80(2), 247-264.

Burla, L., Knierim, B., Barth, J., Liewald, K., Duetz, M., & Abel, T. (2008). From text to codings: Intercoder reliability assessment in qualitative content analysis. Nursing research, 57(2), 113-117.

Cali, D. D. (2002). Journalism after September 11: Unity as moral imperative. Journal of Mass Media Ethics, 17(4), 290-303.

Carey, S. (2002). Undivided loyalties: Is national identity an obstacle to European integration? European Union Politics, 3(4), 387-413.

CBS (2010, January). Nederland langs de Europese meetlat. Retrieved from

http://www.cbs.nl/NR/rdonlyres/ADD6E347-7E97-4564-B0CF-97F16C9DD8F9/0/2010a312pub.pdf.

De Vocht, A. (2013). Basishandboek SPSS 21. Utrecht: Bijleveld Press.

De Vreese, C. H. (2001). Europe in the news a cross-national comparative study of the news coverage of key EU events. European Union Politics, 2(3), 283-307.

De Vreese, C. H. (2007). A spiral of Euroscepticism: The media's fault? Acta Politica, 42(2), 271-286.

De Vreese, C. H., Banducci, S. A., Semetko, H. A., & Boomgaarden, H. G. (2006). The news coverage of the 2004 European Parliamentary election campaign in 25 countries. European Union Politics, 7(4), 477-504.

De Vreese, C. H., & Boomgaarden, H. G. (2005). Projecting EU referendums fear of immigration and support for European integration. European Union Politics, 6(1), 59-82.

De Vries, C. E., & Edwards, E. E. (2009). Taking Europe to its extremes: Extremist parties and public Euroscepticism. Party Politics, 15(1), 5-28.

DellaVigna, S., & Kaplan, E. (2007). The Fox News effect: Media bias and voting. Quarterly Journal of Economics, 122(3), 1187-1234.

Deuze, M. (2003). The web and its journalisms: Considering the consequences of different types of news media online. New Media & Society, 5(2), 203-230.

Eichenberg, R. C., & Dalton, R. J. (1993). Europeans and the European Community: The dynamics of public support for European integration. International organization, 47(4), 507-534.

Elvers, G. (n.d.). Using SPSS for t tests. Retrieved from

http://academic.udayton.edu/gregelvers/psy216/spss/ttests.htm.

Gleissner, M., & De Vreese, C. H. (2005). News about the EU Constitution: Journalistic challenges and media portrayal of the European Union Constitution. Journalism, 6(2), 221-242.

Haas, E. B. (1968). The uniting of Europe: Political, social, and economic forces 1950-57. California: Stanford University Press.

Hallin, D. C., & Mancini, P. (2004). Comparing Media Systems. Three models of media and politics. Chapter 8 “The forces and limits of homogenization”.

Hendriks Vettehen, P. G. J., Nuijten, C. M., & Peeters, A. L. (2008). Explaining effects of sensationalism on liking of television news stories: The role of emotional arousal. Communication Research, 35(3), 319-338.

HOI (2015, January). NOM publiceert oplagecijfers 2014 Q3. Retrieved from http://www.hoi-

online.nl/Doelpagina-Artikellijst/2045/1461026/NOM-publiceert-oplagecijfers-2014-Q3.html.

Hooghe, L., & Marks, G. (2004). Does identity or economic rationality drive public opinion on European integration? Political Science and Politics, 37(3), 415-420.

Hooghe, L., & Marks, G. (2005). Calculation, community and cues public opinion on European integration. European Union Politics, 6(4), 419-443.

Hooghe, L., & Marks, G. (2009). A postfunctionalist theory of European integration: From permissive consensus to constraining dissensus. British Journal of Political Science, 39(1), 1-23.

Krouwel, A., & Abts, K. (2007). Varieties of Euroscepticism and populist mobilization: Transforming attitudes from mild Euroscepticism to harsh Eurocynicism. Acta politica, 42(2), 252-270.

Lubbers, M. (2008). Regarding the Dutch ‘Nee' to the European Constitution: A test of the identity, utilitarian and political approaches to voting ‘No'. European Union Politics, 9(1), 59-86.

Lubbers, M., & Jaspers, E. (2011). A longitudinal study of Euroscepticism in the Netherlands: 2008 versus 1990. European Union Politics, 12(1), 21-40.

Lubbers, M., & Scheepers, P. (2005). Political versus instrumental Euro-scepticism: Mapping scepticism in European countries and regions. European Union Politics, 6(2), 223-242. Lubbers, M., & Scheepers, P. (2010). Divergent trends of Euroscepticism in countries and

regions of the European Union. European Journal of Political Research, 49(6), 787-817.

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.

Maier, S. (2010). All the news fit to post? Comparing news content on the web to newspapers, television, and radio. Journalism & Mass Communication Quarterly, 87(3), 548-562. Manning, P. (2001). News and news sources: a critical introduction. London: Thousand

Oaks.

McElroy, G., & Benoit, K. (2012). Policy positioning in the European Parliament. European Union Politics, 13(1), 150-167.

McLaren, L. M. (2002). Public support for the European Union: Cost/benefit analysis or perceived cultural threat? The Journal of Politics, 64(2), 551-566.

McLaren, L. M. (2007). Explaining mass-level Euroscepticism: Identity, interests, and institutional distrust. Acta Politica, 42(2), 233-251.

Mokken, R. J. (1971). A theory and procedure of scale analysis. The Hague: De Gruyter Mouton.

Sanoma (2013, November). Onderzoek: Nu.nl de grootste nieuwssite. Retrieved from

www.sanoma.nl/pagina/persnieuws/onderzoek-nunl-de-grootste-nieuwssite/.

Schuck, A. R., Xezonakis, G., Elenbaas, M., Banducci, S. A., & De Vreese, C. H. (2011). Party contestation and Europe on the news agenda: The 2009 European Parliamentary Elections. Electoral Studies, 30(1), 41-52.

Semetko, H. A., & Valkenburg P. M. (2000). Framing European politics: A content analysis of press and television news. Journal of Communication, 50(2), 93-109.

Shehata, A. (2010). Marking journalistic independence: Official dominance and the rule of product substitution in Swedish press coverage. European Journal of Communication, 25(2), 123-137.

Steenbergen, M. R., Edwards, E. E., & De Vries, C. E. (2007). Who’s cueing whom? Mass-elite linkages and the future of European integration. European Union Politics, 8(1), 13-35.

Strömbäck, J., & Dimitrova, D. V. (2006). Political and media systems matter: A comparison of election news coverage in Sweden and the United States. The Harvard International Journal of Press/Politics, 11(4), 131-147.

Szczerbiak, A., & Taggart, P. (2003, March). Theorising party-based Euroscepticism:

Problems of definition, measurement and causality (Sussex European Institute Working Paper No. 69). Nashville: Sussex European Institute.

Taggart, P. (1998). A touchstone of dissent: Euroscepticism in contemporary Western European party systems. European Journal of Political Research, 33(3), 363-388.

Takens, J., Ruigrok, N., van Hoof, A. M., & Scholten, O. (2010). Old ties from a new(s) perspective: Diversity in the Dutch press coverage of the 2006 general election campaign. Communications, 35(4), 417-438.

Tewksbury, D. & Rittenberg, J. (2012). News on the internet: Information and citizenship in the 21st century. New York: Oxford University Press.

Van der Laan, S. (2013, January). EU-commissie: Brussel moet journalisten controleren. Retrieved from

http://www.elsevier.nl/Europese-Unie/nieuws/2013/1/EU-commissie-Brussel-moet-journalisten-controleren-1154426W/.

Van Spanje, J., & De Vreese, C. (2014). Europhile media and Eurosceptic voting: Effects of news media coverage on Eurosceptic voting in the 2009 European Parliamentary Elections. Political Communication, 31(2), 325-354.

Waterfield, B. (2013, February). EU to set up euro-election 'troll patrol' to tackle Eurosceptic surge. Retrieved from

http://www.telegraph.co.uk/news/worldnews/europe/eu/9845442/EU-to-set-up-euro-election-troll-patrol-to-tackle-Eurosceptic-surge.html.

Weßels, B. (2007). Discontent and European identity: Three types of Euroscepticism. Acta Politica, 42(2), 287-306.

Williams, P., & Nicholas, D. (1999). The migration of news to the web. Aslib Proceedings, 51(4), 122-134.

APPENDIX A

SAMPLING PROCEDURE

Sampling procedure nu.nl

Time period: September 1st, 2014 – December 1st, 2014. Keywords:

 Europese Unie

 EU

 Europese Commissie

Search results per keyword:

 Europese Unie (N = 25)

 EU (N = 13)

 Europese Commissie (N = 45)

Total results: N = 83

After removing duplicate search results: N = 78. Total EU related articles from medium: N = 78.

Sampling procedure De Telegraaf

Time period: September 1st, 2014 – December 1st, 2014. Keywords:

 Europese Unie

 EU

 Europese Commissie