European Elections and Domestic Politics: Lessons from the Past and Scenarios for the Future

Appendix B The media study

Appendix B The media study edmund lauf and jochen peter Research on television news has shown that the election coverage tends to cluster shortly before Election Day, and European election content is mainly broadcast in the main evening news (see Blumler 1983; Reiser 1994). Rather than short bulletins or current affairs programs, television news during the 1999 European election campaign was therefore analyzed for the two weeks prior to Election Day. When the first European Parliamentary elections were held in 1979, public service television was the norm in each country. By 1999, however, most countries had strong private television channels, and research indicates that there may be consequences for regular viewers’ perceptions of, knowledge about, and willingness to participate in, democracy (Aarts and Semetko 2003). We selected the main evening public and commercial news programs with the highest viewership at the time of the post-election survey in 1999. For Belgium we analyzed Dutch-speaking Flanders and French-speaking Wallonia separately. Given that no private channels existed in Austria or were of no importance in Ireland in 1999, only the public broadcasting channel with the largest reach was included in these two countries. Because only a minority watched the Greek public broadcasting channel, ET1 (Seri 2002), a second private channel was analyzed in Greece. Due to its 277 limited reach in comparison to networks in other countries, the channel in Luxembourg was not part of the analysis. The content analysis was organized centrally at the University of Amsterdam, supported by grants from the Dutch National Science Foundation. The research team in Amsterdam (Semetko, Schönbach, de Vreese, Peter, and Lauf) developed the codebook, drawing upon topic and actor coding used in numerous national election campaign studies (Semetko et al. 1991; Semetko and Schönbach 1994; Diez-Nicolas and Semetko 1999; Semetko 2000; de Vreese, Peter, and Semetko 2001), as well as variables used by Schulz (1983) in his analysis of information in the first European Election campaign in 1979. Research on news frames (Semetko and Valkenburg 2000) and on the effects of news on public opinion about European integration (Semetko, van der Brug, and Valkenburg 2003) contributed to the measures developed to assess the visibility of frames in the news. We focused on election and economic frames. Some broad categories about the program were coded on the outlet level, but the focus was on the news story level. The single news story (defined as semantic entity with at least one topic delimited from another story by a change of topic) was the coding unit, a paper sheet had to be filled per story. Main parts of the codebook (including e.g. framing categories, election categories, and actors) had to be coded only if the story was political. Political stories were defined as follows: in the story, politicians or political groups, institutions, or organizations were verbally mentioned at least twice, or were verbally mentioned once and quoted, or were verbally mentioned at least once and depicted at least once. Because the period of investigation was heavily dominated by stories about the Kosovo war and because the focus of the study lay on the coverage of the European election campaign, Kosovo stories were only coded if they clearly referred to the EU. Up to six actors were coded in each story. In addition, the evaluation of the actor in the story was coded for tone, if any, as well as the function of the actor. The codebook also captured whether actors were cited in the news or quoted. The project was organized bilingually. Codebooks and sheets were solely in Englishlanguage,codertraining,pre-andreliabilitytestingwasdoneusingEnglish language material while coding was done by native speakers. They were trained over six weeks before coding, then there were tests for inter-coder reliability, and they were supervised throughout the entire coding process. For each country, the stories were randomly assigned to the coders. Because in cross-national comparative content analyses, differences between the countries can be the unintended result of lacking coordination of the various country groups, the coder trainers of the country groups were in daily contact to coordinate the coding in the country groups and to resolve problems. Moreover, the majority of the coding was 278 Edmund Lauf and Jochen Peter centrally done at the University of Amsterdam to keep the coding process as comparable as possible.1 For the reliability test, coders of all country groups had to code at least eighteen...