The Use of Party Web Sites and Effects on Voting: The Case of the European Parliamentary Elections in Germany in 2009

ABSTRACT

Content analyses have shown that German political parties make use of party Web sites in different ways, leading to unequal levels of mobilizing, informing, interacting, and participating features. This article examines whether or not the use of party Web sites has an impact on voting by using individual-level data from a representative survey of Germany in the European parliamentary elections of 2009. Although voting behavior was primarily influenced by well-known factors such as party identification and issue competence, evidence suggests that the use of party Web sites could have positive effects on voting for the Greens and the Social Democrats.

Keywords:

• Political communication
• voting
• Web campaigning

Acknowledgments

The author would like to thank Jens Wolling, Jens Vogelgesang, and the anonymous reviewers for their helpful comments.

Notes

1. The data set for replication is provided online at http://dvn.iq.harvard.edu/dvn/dv/jitp. Replication materials are available upon request.

2. The effect of candidate orientation on voting, as posited by the social psychological approach, was not taken into account. This concept does not apply to the European parliamentary elections, where, especially in Germany, most people do not even know the candidates that are running for a parliamentary seat (Niedermayer, 2005).

3. Means and standard deviations are based on the final sample with imputed missing values (see the “Analytical Approach” section, n = 669) consisting of voters and nonvoters excluding cases where the interviewee refused to tell the party whom he or she had voted for.

4. According to German statistical data (Bundesamt für politische Bildung, 2008).

5. The path analyses were run using AMOS, Version 20.

6. Using SPSS, Version 20. EM algorithms are mainly based on iterative regression analyses using the complete values as predictors of the missing values (Graham, 2009). One additional case, which still had a missing value after imputation, was excluded.

7. Due to theoretical reasons, the considered variables in the models differed slightly. Whereas “Religion: Catholic” was included in the CDU/CSU model, the SPD model considered “Religion: Protestant,” and the other more secular-oriented parties' models (Bündnis90/Die Grünen, FDP, Die Linke) included “No religion.”