Computational methods like automated content analysis are increasingly used to gather and analyze data in many social sciences, including journalism studies. Drawing on a systematic literature review on “text as data” approaches in journalism studies, the talk illustrates two points: First, how the method has been used, indicating a strong focus on empirical applications rather than methodological advances. Second, which pitfalls come with the so-called “computational turn”, in particular considering interdisciplinarity in terms of theories, methods, and collaborations within the social sciences.