Unboxing Computational Social Media Research From a Datahermeneutical Perspective: How Do Scholars Address the Tension Between Automation and Interpretation?

Abstract

Communication researchers have fruitfully applied computational methods in their analysis of communication processes. However, the automation of scientific data collection and analysis confronts scholars with fundamental epistemological and practical challenges. Particularly, automation implies that the processing of data is highly standardized for all cases. In the context of social science research, this contrasts with the expectation that meaning is always attributed in individual interaction processes. Based on a literature review of peer-reviewed journal articles, our study explores the resulting tension between automated and interpretive research. We first analyze the extent to which automated methods play a role in social media research. We then identify the challenges and limitations researchers addressed in their studies. On this basis, we propose steps for a data hermeneutical perspective that combines computational methods with interpretive approaches.