In this paper, we explore narratives that emerge in the Twitter discourse during high stakes, low probability events. In particular, we analyze 7.4 million tweets related to four shooting events in the United States of America to identify differences that arise in the semantic structure and message diffusion of mainstream narratives on the one hand and alternative narratives on the other. Our findings indicate that alternative narratives introduce keyterms that have little to no connection to the respective shooting itself and that their diffusion patterns similar to those of mainstream narratives. Moreover, we found empirical evidence of alternative narratives, such as false flag accusations, that appear across different events and persist in the Twitter-sphere over an extensive period of time.

Top