Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Room
Browse By Division
Browse By Session Type
Search Tips
About Seattle
Personal Schedule
Sign In
Session Submission Type: Panel
Journalism has undergone enormous changes in recent years—the rapid development of computer technology and the digitization of content, the impact of online communication, a global economic crisis, and wide-reaching societal and political changes are some of the factors that have considerably altered the face of journalism. It has been argued that journalism research must adapt to the situation and change accordingly, using innovative approaches and methods. Furthermore, the technological changes do not only affect journalism, but also science, offering new possibilities and paths for research.
One of the most debated areas of development is the analysis of ‘Big Data’: A wealth of digital content can be retrieved from websites or news archives, and in its digital form, it is seemingly ready for analysis. The promises of Big Data are tempting: By accessing large chunks of the journalistic output over longer periods of time, researchers can get a topological overview of complete corpora using (semi)automatic analysis, while still retaining the full material for detailed analysis. Furthermore, large-scale comparative approaches to content analysis seem to be much easier to conceive. However, some doubts remain about the feasibility and the adequacy of new computerized analysis under actual research conditions.
In our panel, we will therefore discuss four Big Data projects—their theoretical background, their methodological rationale, the technologies they use, the findings they offer, and the limitations they still have. The corpus size of these studies ranges from 150.000 to 6.7 million content units. The panel will have a deliberate ‘workshop’ character, not only presenting ‘processed’ findings, but also research paths and the pitfalls of Big Data analysis.
The panel will open with an introductory presentation giving an overview of various approaches in the field (e.g., deductive or inductive, semi-automatic analysis vs. fully automatic classification). A typical process of Big Data analysis will be described as a framework, highlighting the most accessible options for journalism researchers from automatic coding using pre-defined dictionaries, named entity recognition, topic modeling, clustering and pattern extraction, to machine learning. Four individual presentations will flesh out that framework with experiences from large-scale projects, focusing on journalistic websites, blogs, Twitter messages, and electronic archives. These projects cover part of the range of what is currently feasible, offering insight into the analysis of hundreds of thousands of news pieces—but they will also uncover the current challenges when researchers undertake projects that involve Big Data analysis.
Introduction: Big Data Content Analysis in (Online) Journalism Research - Thorsten Quandt, University of Muenster
Foreign Nation Visibility in the Associated Press: A Longitudinal Analysis of Contextual Determinants - Rodrigo Zamith, U of Minnesota; Seth C. Lewis, U of Minnesota
Observing the News Flow: A Big Data Analysis of Online Coverage During the 2013 German Elections - Elisabeth Guenther, University of Muenster; Thorsten Quandt, University of Muenster
Big Data Analysis of Public Service Broadcasting: Problems and Solutions in Longitudinal Semiautomatic Analysis of Online News - Eirik Stavelin, U Bergen; Helle Sjovaag, University of Bergen; Hallvard Moe, U of Bergen
Hashtag Dissent: Finding Meaning in the Counter-Narrative of the #Idlenomore Protests in Canada - Alfred Hermida, University of British Columbia; Candis Callison, University of British Columbia