Eunae Yoo, Arizona State University
William Rand, University of Maryland
Mahyar Eftekhar, Arizona State University
Elliot Rabinovich, Arizona State University

The rapid diffusion of information is critical to combat the extreme levels of uncertainty and complexity that surround disaster relief operations. As a means of gathering and sharing information, humanitarian organizations are becoming increasingly reliant on social media platforms based on the Internet. In this paper, we present a field study that examines how effectively information diffuses through social media networks embedded in these platforms. Using a large dataset from Twitter during Hurricane Sandy, we first applied information diffusion theory to characterize diffusion rates. Then, we empirically examined the impact of key elements on information propagation rates on social media. Our results revealed that internal diffusion through social media networks advances at a significantly higher speed than information in these networks coming from external sources. This finding is important because it suggests that social media networks are effective at passing information along during humanitarian crises that require urgent information diffusion. Our results also indicate that dissemination rates depend on the influence of those who originate the information. Moreover, they suggest that information posted earlier during a disaster exhibits a significantly higher speed of diffusion than information that is introduced later during more eventful stages in the disaster. This is because, over time, participation in the diffusion of information declines as more and more communications compete for attention among users.


This is a working paper.