Abstract
This edited volume has come out of the collaborative effort of the team behind the German Longitudinal Election Study (GLES). GLES, Germany’s national election study and a long-term project funded by the Deutsche Forschungsgemeinschaft (DFG), is located at the universities of Mannheim and Frankfurt, the GESIS Leibnitz Institute for the Social Sciences and the WZB Berlin Social Science Centre. As this book was prepared in the wake of the 2013 election, this is the most recent election it covers. Many contributions also make use of data from the project’s first iteration in 2009.
In addition to the standard pre- and post-election cross-section, the GLES also features a rolling cross-section, a campaign panel and a long-term online panel, among others. The individual contributions assembled in this volume make comprehensive usage of this variety of data sources and combine data from multiple sources. Yet the editors do not frame the book as one by or about the GLES. Rather, what binds its diverse contributions together is the theme of context in citizens’ opinion formation and voting behaviour. The editors define context broadly as ‘all features that describe the surroundings in which people live, form political opinions, and make voting decisions’ (p. 4). These are, for instance, parties and candidates (Chapters 5, 8 and 10), media reporting (Chapters 2, 3, 4, 11 and 12) or aggregate unemployment (Chapter 7).
Context can be measured indirectly, as citizens’ perceptions trough specific survey items, or directly. For the latter, some authors make use of electoral returns and socio-economic data at the aggregate level which they combine with survey data in multilevel models (Chapters 6 and 7). It is noteworthy that the GLES also captures context directly through a candidate survey and media content analysis (which are used in Chapters 2, 3, 4, 5, 11 and 12).
The editors motivate the volume’s focus on context by framing context effects as critical evidence for democratic accountability. In doing so, they associate individual-level predictors, interpreting them in a socio-psychological perspective as long-term dispositions, with continuity and context with (short-term) change. Electoral changes can lead to citizen-induced changes of government which are a prerequisite for effective accountability. More specifically, they distinguish distant and proximal contexts. The former are constituted by institutions and social conditions and the latter streams of communication through which citizens learn about these more distant contexts. That theoretical distinction becomes empirically relevant; for instance, in Partheymüller’s contribution on issue salience in which she brings together self-reported media exposure and coding of media content (Chapter 2).
The ordering of contributions in this volume follows a simple temporal logic. While Chapters 2 to 5 focus on the campaign period, the remainder of the book deals with voter behaviour in the election: Chapters 6 to 8 focus on turnout, and Chapters 9 to 12 focus on vote choice. In Chapters 2 and 3, the respective authors investigate the effect of media coverage on issue salience as measured by the most important problem questions. In Chapter 3, Wagner and Werner address the question of whether televised debates lead to issue specific learning on the side of voters. In the following chapter, Weßels focuses on the personalization of electoral competition in district races in Germany’s mixed-member proportional system.
In Chapter 6, Schäfer and Schmitt-Beck test whether individual turnout depends beyond personal characteristics also on aggregate turnout in a voter’s constituency. The authors of the following chapter investigate, in a similar fashion, whether the unemployment rate in a respondent’s place of residence correlates with their turnout decision. Chapter 8 also deals with turnout but focuses on citizens’ perceptions of parties’ ideological distances to themselves and each other.
The GLES online panel also covered thirteen state elections which took place between March 2011 and September 2014. In Chapter 9, Giebler presents identical models of vote choice on all elections in that period showing that voting patterns only rarely differ between the federal state and the national arena.
The authors of Chapter 10 show that a similarity between a candidate and an individual voter makes voters without partisan identification more likely to vote for such a candidate. In Chapter 11, Mader and Schön study the effect of media context on ideological voting. Finally, Chapter 12 looks at the timing of voting decisions in both the 2009 and the 2013 elections. Preißinger shows that more than half of the panel voted at different times in the two elections.
The editors frame this book as, above all, a substantive contribution to the literature on the role of context in voting. And the volume fulfils this purpose well. However, the unique combination of data sources that the individual contributions make use of is of great interest to researchers in and of itself. One could have wished for a more detailed description of the GLES project, be it in a dedicated chapter, in the preface or in the individual contributions. The latter are sometimes short on the details of the data and in particular on their collection.
Furthermore, some contributions strike the reader as unnecessarily complex as they feature a few too many hypotheses. One contribution tested as much as eight different hypotheses. This can make it difficult at times to distil key learnings from the contributions. This proliferation of hypotheses also results in an oversupply of tables, the median being four tables per contribution. Some of them simply provide descriptive statistics which could have easily been visualized for more intuitive communication.
Overall, this volume provides many new insights by combining unique data sources and innovative methodologies. In this regard, it should be noted that replication materials for all contributions are available at https://doi.org/10.17605/OSF.IO/QF29R. Thus, the volume should serve researchers as a highly useful template for their own research, possibly using data on the 2017 election collected by GLES.