Mediated Contestation in Comparative Perspective

Research question/goal: 

Mediated contestation is an important arena for the articulation of identities and interests as well as a crucial context for democratic governance and problem solving. This project aims at identifying the relevant macro-social and media-related preconditions of mediated contestation as well as systematically assessing them from different normative perspectives.

The extent, structure, content and style of mediated contestation over issues related to religion/secularism are analysed in six democracies (USA, Australia, Germany, Switzerland, Turkey, and Lebanon) and three media types (daily newspapers, news websites, and political blogs). The project tests hypotheses regarding the influence of two macro-social conditions and two important media attributes. The two macro conditions are (1) the structure of the political system (majoritarian vs. consensus democracies) and (2) the existence or non-existence of a deep cultural division (contested vs. uncontested secularism). The media attributes studied are (3) the degree of users’ opportunities to respond to media content (low for daily newspapers vs. high for news websites and political blogs) and (4) the level of opinion orientation (low for daily newspapers and news websites vs. high for political blogs). In the first part of the project representative and comparable samples of media material will be analysed using standardized content analysis as well as automated topic modeling. Data analysis will rely on multilevel regression models. A follow-up study will later be proposed for continued funding, in which a series of comparative case studies will be conducted following the logic of Lieberman’s nested analysis. These case studies will be based on extended media samples (including social media) and shed light on the causal mechanisms that underlie the formation and characteristics of mediated contestation. In a final step these empirical patterns are subjected to a multiperspectival normative assessment that uses standards derived from liberal, republican, deliberative, and agonistic theories of democracy.

Current stage: 

We are currently analysing the project’s raw data. The collection of digital and non-digital text data encompasses about 1.6 million text documents, published between August 2015 and July 2016 in 119 newspapers, news websites, and blogs from six countries (Germany, Switzerland, Turkey, the USA, Australia, and Lebanon). In a first step, data were preclassified by topic using a new method developed in collaboration with the Data and Web Science Group (DWSG) in the School of Business Informatics and Mathematics. This method (Keyword-Ranked Topic Modeling) combines data from expert surveys in the six countries under study with automated text classification methods. In a second step, the project team collaborated with the DWSG to develop an online tool (MedCon Relevance Coder) for the efficient manual classification of preclassified text documents by a multilingual coder team. This software also offers extensive functionalities for process control and quality management to researcher and will be made freely available to the scientific community. In a third step, we are currently performing final coding of the thematically preclassified text data. For this purpose, the multilingual coder team in Mannheim has been trained to work with a software tool for the manual coding of text data (Angrist), which has been adapted to the needs of the project in collaboration with partners at the University of Zurich. Previous theoretical and methodological project findings have been presented at several conferences and are being prepared for publication in top-tier academic journals.

Fact sheet

2012 to 2020
Data Sources: 
Quantitative media content data (own data gathering)
Geographic Space: 
USA, Australia, Germany, Switzerland, Turkey, Lebanon