POLITICAL COMMUNICATION ABOUT EUROPE ON THE INTERNET DURING THE 2004 EUROPEAN PARLIAMENT ELECTION CAMPAIGN IN NINE EU MEMBER STATES

It is generally acknowledged that public support for European Union (EU) policy and institutions is low in most member states; European Parliament (EP) elections have been frequently criticized as having little significance to voters: since they are – still – organized along national lines instead of European ones, they cannot be considered truly ‘European’ elections, but rather mid-term national contests, or ‘second order national elections’, which focus on domestic rather than European issues. As a result, European issues and events tend to attract less attention from all political actors such as political parties, interest groups and the mass media in comparison to national political issues (Franklin 2001; Hix 2005: 177, 193).

It is within this context that scholars have begun to recognize that the process of European integration from above must be accompanied by Europeanization of political communication in order to overcome the lack of legitimacy and popular involvement in the EU by European citizens (Koopmans et al. 2000: 12; Koopmans and Pfetsch 2003). Political communication about Europe, facilitated within a European public sphere, could reduce the so-called ‘democratic deficit’ by enhancing processes of identification among participants living in different EU member states (e.g., Kunelius and Sparks 2001; Koopmans et al. 2004).

In this article, we explore the manner in which the concept of ‘European public sphere’ has been formulated during the last decade, and present a concrete approach to investigating this concept with direct reference to the Internet. Specifically, as an indicator of Europeanization of political communication, and subsequently of the existence of a European public sphere, we examine the visibility of communication about Europe on websites produced by a variety of political actors in nine European countries. This study has been performed in the context of the 2004 EP election in the following EU member states: Czech Republic, Finland, France, Hungary, Ireland, Italy, The Netherlands, Slovenia and the United Kingdom. The investigation is guided by the following research question: To what extent do national political actors address European issues on their websites in the context of the 2004 EP election campaign? This article stems from a pan-European collaborative empirical investigation exploring the use of the Internet during the 2004 EP election and analysis of data collected by research teams in nine EU member states.¹

In the current discussion on European integration, the notion of the public sphere, initially elaborated by Jürgen Habermas, has begun to play a central role. This notion, as elaborated by Habermas (1962/1989), places emphasis on the deliberative and discursive aspects of democracy. Given that public affairs increasingly extend beyond the boundaries of the nation state, scholars have argued for a new public sphere functioning at the European level (e.g., Kunelius and Sparks 2001). However, in this context scholars seem to disagree about whether a supranational European public sphere is emerging, or whether one should merely speak of ‘Europeanization’ of national public spheres when referring to Europeanization of political communication. Early scholars dealing with the possibility of a public sphere functioning at the European level retain the original Habermasian notion of the public sphere. This notion, generally speaking, involves ‘the space within which the affairs of the state could be subjected to public scrutiny’ (Kunelius and Sparks 2001: 11). These scholars argue that a ‘genuine’ supranational European public sphere can only emerge once Brussels becomes, at least partially, an independent political centre with its ‘own’ political actors, its ‘own’ mass media, and its ‘own’ public – which is not the case in the present situation in which the EU is faced with a democratic deficit (e.g., Gerhards 1993, Grimm 1995; Schlesinger 1996, 1999; Schlesinger and Kevin 2000). In various degrees, they place emphasis on the lack of transnational political actors like political parties and interest groups at the European level, the lack of European-level mass media, the diversity of languages across Europe, and the absence of a collective identity; for this reason, most of these scholars prefer to speak of ‘Europeanization of national public spheres’ when investigating Europeanization of political communication.

Other, more recent scholars, consider this view about what constitutes a European public sphere as too restrictive, and based on an idealized picture of an almost homogeneous national public sphere, which is then replicated at the European level (e.g., Risse and Van de Steeg 2003). According to Koopmans and colleagues, for example, this view ‘presupposes a degree of linguistic and cultural homogeneity and political centralisation that cannot be found in many well functioning democratic nation-states’ (Koopmans et al. 2004: 10). Koopmans et al. (2000) use Switzerland as an example of a country that has managed to create a national public sphere despite the presence of three language groups and no newspapers that can be considered national in character. Instead, these and other scholars place emphasis on parallel public debates across Europe as an indicator of the existence of a European public sphere. In this view, a European public sphere must be considered a social construction produced through discursive practices (Eder et al. 2000; Risse 2002, 2003; Van de Steeg 2002 Van de Steeg 2004; Risse and Van de Steeg 2003). As Risse (2003: 4) argues; ‘A European public sphere does not fall from heaven, and does not pre-exist outside social and political discourse. Rather, it is being constructed through social and discursive practices creating a common horizon of reference and, at the same time, a transnational community of communication over issues that concern ‘us as Europeans’ rather than British, French, Germans or Dutch’.

A rule of thumb often used as point of departure by scholars attempting to empirically measure Europeanization of political communication is offered by Eder and Kantner (2000). They suggest that the key indicator for a shared public debate at the European level is whether similar European issues are being addressed in different national media at the same time, using the same criteria of relevance (Eder and Kantner 2000: 102). Gerhards (2000) extends this proposal by advocating a more normatively demanding stance towards what constitutes Europeanization of political communication. He argues that in order for Europeanization to take place, an actor communicating about a European issue or event, should also ‘evaluate it from a perspective that extends beyond one's country and interest’ (Gerhards 2000: 293). Others seem to agree with Gerhards on this point and consider it of equal importance that actors not only communicate about European issues, but that they also communicate from a European perspective. Related to these different interpretations of what constitutes ‘Europeanization’, roughly two approaches can be separated in measuring Europeanization of political communication.

The first approach essentially counts how often European issues are mentioned in the mass media, thereby measuring the visibility of communication about Europe (Hodess 1997; Gerhards 2000; Kevin 2001; Groothues 2004). For example, Groothues compares the number of (prime-time) news items dealing with EU affairs to (1) news items dealing with other European countries, (2) domestic news items, and (3) non-European items, during two ‘routine’ weeks in 2003 for three television stations located in France (France 2), Germany (ARD) and the UK (BBC 1). For all three stations, only a small percentage, 2–4 per cent, of news items dealt with strict EU affairs. On average, 25 per cent of the news items dealt with events/issues in other European countries (variation between the stations: 20 per cent for France 2, 28 per cent for the BBC, 31 per cent for ARD); a majority of the news items, however, dealt with purely domestic issues (an average of 68 per cent, with differences between the three stations ranging from 57 per cent for ARD, 67 per cent for the BBC and 75 per cent for France 2) (Groothues 2004: 9).

The second more qualitative approach concentrates on analysing media reporting on particular European issues, and focuses on the simultaneous appearance of interpretative patterns, or alternatively ‘frames’, in which European issues are addressed across national media (Risse and Van de Steeg 2003). For example, Semetko et al. (2000) have investigated the extent to which European issues, problems, events and persons in national news are framed as ‘European’ or ‘domestic’. They conclude that European and Brussels-based news has become more important in the last few years for national news media (Semetko et al. 2000: 129). Also, Eder et al. (2000) identified three frames: interests, identity and values and investigated whether these three interpretations of Europe are shared across European countries within the different national media (Eder et al. 2000; see also Trenz 2004: 308–9). In an analysis of news coverage of European governance and policy making during the year 2000, 85 per cent of the articles in the sample contained an interests frame, 37 per cent were coded in normative terms (values frame), and only 27 per cent contained an identity frame (Trenz 2004: 309–10).

In this article, we focus on communication about Europe present on the Internet. We believe that the proposed conceptual frameworks for investigating Europeanization of political communication in the mass media can be also used in the investigation of online communication about Europe on political websites. This article utilizes the first approach, which measures Europeanization of political communication via the visibility of communication about Europe.

The study focuses on the extent to which a variety of political actors such as political parties, NGOs, governmental organizations and press actors address European issues and events on their websites in the context of the 2004 EP election. Similar to Koopmans and Pfetsch (2003), we consider the mass media not only as conveyors of information or channels of communication through which other political actors communicate with the public, but also as political actors (press organizations) in the public sphere who legitimately make their own voice heard, and who nowadays produce their own websites. The investigation of websites makes it possible to treat each different type of actor as an equal participant in the public sphere.

The Internet, it is often claimed, potentially provides a space that is accessible to a wide variety of political actors – citizens, governments, political parties, advocacy groups and alternative social movement organizations – to share information, discuss issues, and propose and engage in political action, on- and offline (e.g., Mitra and Cohen 1999: 180). These elements are often considered important components of the political process and accordingly the public sphere (Tsagarousianou 1999; Jankowski and Van Selm 2000). Scholars, however, disagree about the impact of such engagement. First-generation ‘cyber-optimists’ stress the opportunities for deliberation and direct decision-making among a broad spectrum of the public in an Internet environment (e.g., Rheingold 1993; Rash 1997). Later, ‘cyber-pessimists’ warn that the Internet may widen the gap between the engaged and the apathetic (e.g., Margolis and Resnick 2000). These scholars claim that cyberspace increasingly reflects the political forces that dominate politics and social life in the real world, and that ‘political life on the Net is therefore mostly an extension of political off the Net’ (Margolis and Resnick 2000: 2–3, 14).

Norris suggests that a balance should be found between these two extremes and proposes a middle ground position, thereby positioning herself as a ‘cyber-skeptic’ (Norris 2001: 233–9). She emphasizes the possibility for substantial transformation of the political arena when ‘transnational advocacy networks and alternative social movements … have adapted the resources of new technologies to communicate, organise, and mobilise global coalitions around issues’ (Norris 2001: 238–9). Similarly, Foot and Schneider (2002) argue that the impact of the Internet can be found in changes at the structural level of the political system. They stress the importance of independent political websites developed by national and state advocacy groups, civic organisations and the mainstream and alternative press. Complementing this perspective, they view the websites of political parties as components within a larger political arena (Foot and Schneider 2002).

Ward et al. (2003a) summarize the main areas where political transformation has been anticipated and argue that, although a revolutionary transformation of politics is not to be expected, ‘our early research indicates that the Internet will make a modest positive contribution to participation and mobilisation’ (Ward et al. 2003b: 667). In all likelihood, as others (Bimber 1998) have suggested, the Internet may play a role in conjunction with other societal trends in contributing to a transformation of politics.

These ideas and studies constitute the backdrop for our investigation of the use of the Internet by political actors during the 2004 EP election campaign. During recent years more and more websites, produced by a variety of political actors, have become available to citizens for political communication on European issues and events. In the case of the European Parliament elections, often considered ‘second order national contests’, it is important for political actors to maintain websites as a means of communication with the electorate for several reasons. First, European issues and events are generally less intensely covered by the mass media than national political issues (Thomassen and Schmitt 1997; Hix, 2005: 193) and therefore, websites may be an effective alternative for disseminating European issues. Second, during the last two decades there has been a tendency for the media to offer analysis rather than straight of reporting of election campaigns. As Gulati et al. (2004: 243) stress, it has become acceptable for both television and the print media to include ‘an interpretative, and therefore inherently subjective, component to their campaign coverage’. The Internet seems to be the perfect medium for political actors to circumvent this ‘interpretative reporting’, and to inform citizens directly and systematically about their positions on EU policy and legislation for the upcoming period (see also Nixon and Johansson 1999; Ward et al. 2003a).

For the reasons outlined above, namely the ‘second order’ image of EP elections and the ‘democratic deficit’ of the European Parliament in general, one may question whether an EP election serves as appropriate occasion to measure the existence of a European public sphere (see also Van de Steeg 2004: 145). On the other hand, previous studies have supported the notion that media attention to EU events and issues is cyclical, and can mainly be found in the context of major EU events such as EP elections (De Vreese 2001). Political actors may apply a similar strategy when addressing political issues on their Web sites. Furthermore, as De Vreese et al. (2001) outline, using a common European event to investigate the mass media reporting on Europe has the following advantage: we may obtain better measures for cross-national comparison of communication about a common event than would be possible with general political and economic coverage about unrelated domestic events.

Here, we acknowledge the importance of political actors addressing European issues and events on their websites, and consider this an indicator of Europeanization of political communication, and, subsequently, of the existence of a European public sphere (see also Zimmermann and Koopmans 2003). In this article we investigate the visibility of European issues and events on websites of political actors in the context of the EP election as played out in 9 EU member states in June 2004, as will be outlined in the next section.

This study is part of an international collaborative investigation concerned with the role of the Internet during election campaigns.² Building on the experience and methodological procedures and tools developed by WebArchivist (http://webarchivist.org/ie), empirical projects were established around national elections in seven Asian countries, the United States, and the 2004 European Parliament election held in 11 EU member states. This study focuses on nine of these 11 EU countries: the Czech Republic, Finland, France, Hungary, Ireland, Italy, The Netherlands, Slovenia, and the United Kingdom. Three of these countries – the Czech Republic, Hungary and Slovenia – are new EU member states as of 2004; five countries – France, Ireland, Italy, The Netherlands and the United Kingdom – have been represented in the EP since the first election in 1979; and Finland first participated in European elections in 1999. The sample cannot be considered as representative for the entire EU, but as a theoretical sample to facilitate development of theory. In the analysis comparisons are drawn between the three ‘new’ Eastern European member states and the six ‘old’ member states (including Finland), and between different actor types. Only websites produced by national political actors are included in the study.

For each country included in the study, during a three-week period running eight to five weeks prior to the election, coders searched for sites they expected to be involved in the 2004 EP election campaign by consulting search engines, politically oriented portals and other depositories of potential website addresses. Much variation exists between the number of sites identified per country, ranging from 123 (Hungary) to 617 (United Kingdom). In countries in which the EP election was organized at the regional level more election-related websites were identified than in countries in which the election was organized at the national level: 318 in France and 617 in United Kingdom. In the Eastern European member states fewer sites were identified than in the older member states: 166 for Czech Republic and 163 for Slovenia, in comparison to 240 for Italy and 318 for The Netherlands.

Five weeks prior to the election, a sample of 100 sites was randomly drawn for each country, stratified for actor types: 30 per cent candidates, 15 per cent political parties, 10 per cent governmental sites, 10 per cent NGOs, 10 per cent labour unions, and the remainder of the sample was distributed across other actor types.³ In the two weeks prior to the election, sites were coded for the presence of 36 information and engagement features. This article focuses on four of these features in order to measure the visibility of communication about Europe on the Internet, which can be considered indicative for Europeanization of political communication: ‘EP election content on front page’, ‘European content on the front page’, ‘European content within the news section’, and ‘European content elsewhere on the site – within two links from front page’. EP election content is defined as ‘website content related to the European Parliament election’; European content is defined as ‘website content related to the EU and/or to the EP election’. Coding was completed before the election took place.

At the time of the actual coding, however, not all content on the 100 sites in the samples could be coded on all variables. Some websites were shut down completely, others were partly dysfunctional. Furthermore, some content was inconsistently coded. Of the 860 initial websites, 18 have been coded ‘missing’ and five sites were not coded on ‘EP election content on front page’ and were subsequently excluded from the study. The total N of the study is therefore 837. Sites that were coded inconsistently on ‘European content on front page’ in comparison to the variable ‘EP election content on front page’ were also excluded from further analysis (110 sites, which lowers the N from 837 to N=727 for this feature). Furthermore, only when websites contained ‘European content on front page’, was content coded on ‘European content within news section’ (N=530), and ‘European content elsewhere on the site’ (N=530).

The analysis took place through cross tabulation of the variable actor type against the variables ‘EP election content on front page’, ‘European content on front page’, ‘European content within news section’ and ‘European content elsewhere on the site’. These two-way cross tabulations were extended by the variable country as an additional dimension. Due to the low number of cases in this three-way cross-tabulation, percentages should be viewed with caution. Since the number of valid cases for variables differs, pairwise deletion of missing cases was applied in the cross tabulations.

As shown in Table 1, the first striking result is the low number of political actors having EP election content on the site front page: 68 per cent. Apparently, for one-third of the actors included in the study – all selected because of their politically active focus or attitude – this event was insufficiently important to consider placing information on their sites. Somewhat more often, however, actors communicated about general European issues on the website front page: 73 per cent. This section reports on observed variations (1) between the countries included in the study, and (2) between the different actor types.

Table 1 reports on the observed variations between the nine EU member states (average percentages calculated per country). For the first variable; ‘EP election on front page’, most countries scored around the average percentage (±10 percentage points). However, both Slovenia and the United Kingdom scored exceptionally high on this feature: respectively, 97 and 92 per cent. These two countries also scored high on the other three variables. Countries scoring much lower on ‘EP election content on front page’ were France and Ireland: respectively, 52 and 31 per cent. France, however, scored much higher on the other variables, 77 per cent for European content on front page, 40 per cent for EP/EU-related news, and 71 per cent European content within two links from front page. For French actors, general European issues were clearly more important than the EP election as a specific event; their online political communication still was quite Europeanized. This was much less the case for the Irish actors, who also scored low on more general European content variables, such as 47 per cent for European content on front page. When comparing the percentages of the three new Eastern European countries with those of the six old member states, no clear pattern can be observed.

Looking more specifically to the observed differences between the various actor types (percentages aggregated for the EU-9 sample), as shown in Table 2, it is especially NGOs and labour unions that did not often provide EP election content on the site front pages (38 per cent), while at the same time these actors scored much higher on the other variables: 46 per cent for ‘European content on front page’, 56 per cent for ‘EP/EU-related news’, and even 84 per cent for ‘European content within two links from front page’. As mentioned, these actors apparently considered ‘Europe’ in general more important as topic for reports than the EP election.

An explanation for this difference in perceived importance by NGOs and labour unions may be found in the long-standing tainted reputation of the EP regarding legitimacy: actors consider the EP as having limited powers in relation to other EU bodies, such as the Council of Ministers. As a result, intense campaigning was not observed outside the institutionalized, traditional actor types of parties and candidates. In comparison, candidates scored 89 per cent for EP election on front page and 95 per cent for general European content on front page; see Table 2. Parties scored respectively, 88 and 90 per cent.

In a more general sense, it is remarkable that in the EU-9 sample the overall percentage for ‘European content within news section’ (53 per cent) is lower than for ‘European content on front page’. Reasons may lie in the level of sophistication of the sites. As elaborated elsewhere (Van Os et al. 2007), websites produced by candidates, governments and labour unions can be quite basic in terms of information and engagement features incorporated, and as a result, often do not have a news section. Last, in relation to ‘European content within 2 links from front page’, it is a well-appreciated practice of press actors and NGOs/labour unions to compose online dossiers focusing on central themes, such as the European Union, or more specifically, the EP election. This practice may have led to a high score on ‘European content within 2 links from front page’ for press actors (85 per cent) and NGOs/labour unions (84 per cent).

Table 3 through 7 elaborate on the EP election/EU related content provided by respective actor types within different sections of their websites, sorted per EU member state. First, when looking at Table 3, candidates provide much content related to both the EP election and to the EU in general: respectively, 89 and 95 per cent. French and Hungarian candidates seem to be the exception to this generally high score: only 64 per cent of the French candidates and 44 per cent of the Hungarian candidates provided EP election content. All French candidates, however, provided general EU content on the front page, in contrast to Hungarian candidates of which only 38 per cent provided general European content on the site front pages. With regard to France, this is probably related to the regional organization of EP elections in France, where political parties established websites for each of the eight regions in the country. Much information was provided on candidates on these sites, which may have lowered the need for active personal campaigning for French candidates in the EP election campaign. Accordingly, these sites provided general European content, but no specific EP election content. No clear explanation is evident for the low level (38 per cent) of general European content on sites of Hungarian candidates. On comparing the percentages of the three new Eastern European countries with those of the six old member states, once again no clear pattern is apparent with regard to the extent to which candidates address EP/European issues.

With regard to Table 4, the most remarkable contribution seems to be the high percentage of Slovenian and British governmental sites providing EP election content on the front page: all governmental sites in these samples provided this type of content. For Slovenia, this may be due to this being its first participation in this pan-European event; governmental organizations may have placed much information, relatively speaking, on the EP election for this reason. However, another eastern European country participating for the first time in this election was Hungary. In this country only 52 per cent of the governmental websites provided EP election content on the front page. The score for European content on the front page was slightly higher in the Hungarian sample: 69 per cent of the governmental sites contained this type of content, about the EU-9 average. In the United Kingdom regional elections were held simultaneously with the 2004 EP election. British governmental actors may have been more motivated to provide content on their websites related to elections in general – the EP election just being one of the two – in comparison with other countries in the study. Furthermore, little European content could be found in the news sections of French and Finnish governmental sites (respectively, 38 and 29 per cent), which is not consistent with the relatively high percentage of European content on front page in the same samples. Similarly, the British governmental sites provided no European content in the news section. As mentioned earlier, governmental sites often do not incorporate sophisticated features such as an up-to-date news section on their websites. This may explain these low percentages. Irish governmental websites scored quite low on the variable EP election on their front pages: 19 per cent. Again, in comparing the percentages of the three new Eastern European countries with those of the six old member states, no clear pattern is present regarding the extent to which governmental organizations address EP/European issues.

Table 5 focuses on websites produced by NGOs and labour unions. Again, Slovenia and the United Kingdom are the positive exception: all six Slovenian and 21 British sites provided both EP election content and general European content on their front pages. As mentioned before, it is especially within this category that large differences can be observed between the variable ‘EP election content on the front page’, and the variable ‘European content on front page’. In particular, the Czech, Finnish and Hungarian NGO and labour union websites contained relatively large amounts of European content on their front pages in comparison with EP election content: the Czech sample (respectively, 61 and 28 per cent), the Finnish sample (respectively, 56 and 28 per cent) and the Hungarian sample (respectively, 59 and 23 per cent). In the samples for these countries, percentages were similar or even higher for the variables ‘European content in the news section’ and/or on a page within two links from the front page.

Again, no substantial differences were found between Eastern and Western European NGO/labour union websites. The general picture for these actors, a low percentage of EP election content and a higher percentage for general EU content, seems to be consistent with the frequently made argument that NGOs and labour unions do not have direct interest in elections. The role of these actors seems to be directed towards lobbying and organizing public actions outside electoral periods.

Table 6 reports on the websites of political parties. Political party websites score relatively high on all variables across the EU-9 sample. It should be mentioned that this category consists of a variety of websites produced by political parties; in addition to a general website and/or a website devoted to the EP election, some political parties also produced their own online journals or websites devoted to youth audiences. These websites did not always contain European content or focus on the 2004 EP election. Furthermore, some websites were produced by political parties that only participate in national elections. In some national samples, as a result, political party websites scored quite low on this variable, especially in the French sample (57 per cent). Dutch, Italian, Slovenian and British party websites, on the other hand, scored 100 per cent on the variable ‘EP election content’, and on ‘European content on front page’. French political parties scored higher on this last variable (86 per cent); they apparently considered ‘Europe’ as a general issue important enough to report on, despite their lack of involvement in the EP election. In contrast, only 33 per cent of the Slovenian political party websites provided EU content within the news section; Finnish political party websites also scored low in this variable (36 per cent). In comparing the percentages of the three new Eastern European countries with those of the six old member states, no clear pattern is evident regarding the extent to which political parties address EP/European issues.

In Table 7 the results are outlined for the actor type press. It should be mentioned again that a press organization was defined as ‘news/publishing organization that creates its own content’. This includes, for example, community portals that are operated by publishers of local magazines and newspapers. Of the press organizations in the EU-9 sample, considered by many as the ‘watchdog’ of state power, only 64 per cent provided content related to the EP election. Large variations could be observed, however, between the countries: only 11 per cent of all the press sites in the Irish sample contained EP election content on their front pages, in contrast to 100 per cent in the Czech sample. However, a substantially higher percentage of the Irish press websites provided general European content elsewhere, either in the news section, or within two links from the front pages. Similar observations were made for other countries. The Eastern European countries in the EU-9 sample seem to score relatively high (Czech Republic 100 per cent, Hungary 86 per cent, Slovenia 89 per cent on the variable EP election) in comparison to the Western European countries. A possible explanation for this difference may be that the press takes it role as ‘watchdog’ of the government in these former Communist states more seriously than is done by press organizations in the Western European countries.

This study suggests that within the context of the 2004 EP election campaign, in nine EU member states, there are multiple websites communicating about Europe and/or the European Parliament election as a specific event. This observation is similar to what Eder and Kantner (2000: 102) define as the key indicator of a shared European public sphere: whether the same European issues are being addressed on various websites in different European countries at the same moment in time.

In one sense, the percentage of websites in the EU-9 sample that contained EP election content, 68 per cent, is not high, in particular because of the search strategy followed. This seems surprising, mainly because of the status of the European Parliament as being one of the two legislative bodies of the EU whose decisions affect all EU citizens. At the same time, the political actors in the sample did communicate more often about general European issues, either on the site front page or on a page within two links from that page. In particular, non-institutionalized and less traditional actors such as NGOs and labour unions used their websites to communicate relatively more frequently about Europe and general European issues than was done about the EP election – apparently these actors considered their own influence in this election minor, or simply did not care about the event because of their own national focus. Still, a high degree of Europeanization could be found on websites produced by these actors. We consider this an indicator of the existence a European public sphere.

The extent of Europeanization also varied substantially between the countries included in the study. Websites produced by British and Slovenian actors often contained communication about Europe, on both the EP election and more general European issues; Irish actors scored relatively low on both variables. Most other countries scored around 68 per cent on the variable EP election. When comparing Eastern and Western European countries, clear variation was only found with the site actor type ‘press’. Individual countries, however, do suggest interesting deviant situations, like the Slovenian governmental organizations and NGOs/labour unions that provided a relatively high amount of European content on their websites. Also, French candidates and parties scored relatively low on the variable EP election content in comparison to the high average scores for the EU-9 sample. These political actors, however, communicated in a more general sense about Europe and European issues on their websites, like most other political actors in the EU-9 sample. In so doing, these individuals and organizations may have influenced the ongoing political discussion during the period of the campaign. In all likelihood, these websites contributed to a general European public sphere: they gained a degree of visibility unavailable through the traditional mass media and made it possible to address particular European issues and express opinions.

This investigation, it should be stressed, is exploratory and, consequently, limited. The fact that political actors simultaneously address particular European issues and events on their websites is only one indicator of the existence of a European public sphere. We know little about the characteristics of the ‘European content’, we only know the extent to which this content is visible on websites of political actors. As outlined in the theoretical section of this article, we consider it of equal importance in the construction of a European public sphere that this online communication contains a European perspective or shared interpretative context. Further research should focus on this aspect. For the time being, however, these results suggest that political actors do use their websites to communicate about European issues and events.

Renée van Os is a Ph.D. candidate at the Radboud University Nijmegen, Department of Communication, the Netherlands. She completed her MA in 2002. Her Ph.D. project focuses on the potential of the Internet for enhancing the European democratic process and contributing to a European public sphere. Van Os is author/co-author of journal articles and book chapters, and has contributed to conference proceedings in which she has investigated aspects of the concept of European public sphere in an online environment. A recent article on the European Parliament election has been published in Javnost-The public.