Social Positions and Collective Sense-Making in Crisis Communication

Abstract

ABSTRACT This article addresses how uncertainties during crisis situations evolve over time and how social positions dynamically affect the collective sense-making process in social media crisis communication. We carried out two case studies on Twitter: (1) the Brussels attacks (2016) with 4,390,784 tweets and (2) the Munich rampage (2016) with 1,258,227 tweets. By applying computed regression-based time-series analyses, we revealed the underlying tweet behavior. As next steps, we trained a machine learning algorithm to identify tweets that express uncertainty and we conducted social network analyses to determine the most influential actors and their social positions. The results reveal that tweet behavior in early crisis stages is dominated by information distribution and guided by content that is characterized by a high percentage of tweets expressing uncertainty. Based on our results, we identified two forms of collective sense-making: (1) acute and guided collective sense-making and (2) evaluative and retrospective collective sense-making.