Brown Bag Research talk with Prof. Kelvin King

Please join Kelvin King as he presents his virtual brown bag talk: Interaction Between Temporal, Content and Creator Characteristics on the Evolution of Viral False and Correction Messages.

Studies indicate that falsehoods are more prevalent when “information during important events are either lacking or subjectively ambiguous”. Traditionally, these events are termed as the “3Cs” – conflict events, crisis scenarios, and catastrophes. According to the Social Media Working Group (SMWG) for disaster management, “one of the biggest challenges public safety agencies and organizations face is how to reduce or eliminate the spread of false information…”. In recent years online social networks (OSNs) have become arteries for misinformation, diminishing their value as effective platforms for disaster management. The proliferation of falsehoods during disasters, saps and undermines the very fabric of our democracy, breeds animosity, erodes public trust, and promotes negative behavior. It is therefore important that we investigate this phenomenon within the context of disaster management.

Despite the burgeoning literature on information diffusion and misinformation, Information Systems (IS) researchers have largely ignored a major aspect of the diffusion process: the morphing of the textual content. In this study, we leverage a big-data theory building approach by Kar & Dwivedi, augmented by machine learning and econometric models in investigating the morphing behavior of more than eighty verified falsehoods and their corresponding correction messages that propagated during 5 major disasters. We fill the gap in disaster management and IS literature that has investigated information diffusion as a static communication process. Investigating how the textual contents of these messages mutate over time, helps us better understand drivers of the propagation of falsehoods during disasters. We identify novel temporal- , creator-, content-characteristics and their interaction terms, that are important predictors of virality and morphing. Furthermore, we demonstrate that “weights” have more predictive power than cascades in predicting virality, morphing behavior and the degree of intensity. Our empirical results further suggests that the classification of falsehoods undergo mutation as they propagate through the ecosystem. We also observe the effects of policy announcements and irrational social behavior and how they catalyze and drive the evolution of these messages on social networks. Finally, we provide valuable insights to practitioners, government agencies, administrators, and policy makers in the development of appropriate mitigating strategies, tactics in engaging the public, the managing of disaster information sharing and in the building of more robust virtual disaster response commands.

This event is free and open to the public. No registration is required.

Join on Zoom