Analysis of discourses in social media

Interdisciplinary Research Project

Period May 2012 to April 2015
Conductor Prof. Dr. Thorsten Quandt (University of Münster)
Prof. Dr. Stefan Stieglitz (University of Münster)
Prof. Dr. Christoph Neuberger (University of Munich)
Prof. Dr. Manfred Stede (University of Potsdam)
Elisabeth Günther M.Sc. (University of Münster)
Institutions Department of Communication (University of Münster)
Department of Information Systems (University of Münster)
Department of Communication Science and Media Research (University of Munich)
Department of Linguistics (University of Potsdam)
Funding Federal Ministry of Education and Research (BMBF)
Tags Social Media, Gatekeeper Research, Discourse Analysis, Methods of Communication Studies


The opportunities of public communication have been extended by social media applications such as Twitter, Facebook, and blogs. On the internet journalistic “gatekeepers” are not central mediators of topics and opinions anymore.

In principle, individual and collective actors of all kinds can participate in public discourses as speakers. Hence, processes and structures of the public are subject to change. Issues and the formation of opinions take different courses than in classical media. At the same time the internet offers a higher degree of transparency. Thus, vast amounts of digital texts can be evaluated automatically.

The goal of the project network is to develop an automated analysis process and evaluate it. The process should be implemented in the form of tools for the analysis of online discourses.


The quantitative content analysis is a manually-conducted standard research method from the field of communication studies that requires vast amounts of both time and money. This is no less the case with qualitative methods used in discourse analysis, which are also used in other social sciences. In order to approach this limitation, the two communication-oriented partners cooperate with representatives from two different research fields, which introduce various methodical approaches.

Research in the field of information systems allows the collection of large amounts of data from different social media applications (data tracking) and is able to structure these data via automated network analysis.

Computer linguistics provides methods for the analysis of sentiments and discourse qualities. Furthermore, an automated method for content analysis based on inductive text classification is to be refined. Also, manual methods will be utilized for evaluative purposes.

Altogether, the project seeks to make versatile and comparative analyses of public discourses on the micro, meso, and macro levels possible.