Word Counts and Topic Models

With digital journalism and social media producing huge amounts of digital content every day, journalism scholars are faced with new challenges to describe and analyze the wealth of information. Borrowing sophisticated tools and resources from computer science and computational linguistics, journalism scholars have started to gain insights into the constant information flow and made big data a regular feature of the scientific debate. Both deductive (manual and semi-automated) and inductive (fully automated) text analysis methods are part of this new toolset. In order to make the automated research process more tangible and provide an insight into the options available, we provide a roadmap of common (semi-)automated options for text analysis. We describe the assumptions and workflows of rule-based approaches, dictionaries, supervised machine learning, document clustering, and topic models. We show that automated methods have different strengths that provide different opportunities, enriching—but not replacing—the range of manual content analysis methods.