As social media has become more present in people’s day-to-day lives, many turn to these platforms in natural disasters to keep abreast of the ever-evolving crisis situation. Facing the increasing amount of crisis-related information available on the social media platforms, users tend to focus on and reach out to authoritative sources—individuals or organizations that provide authoritative and credible crisis-related information due to their official status or position. As they provide high-quality information, response from the authoritative sources can be especially valuable to social media users directly affected in natural disasters. In this study, we aim to extend the reach of credible information during crisis, and direct the attention of authoritative users to the affected users who need their help. Specifically, we investigate what factors differentiate the tweets by regular users that receive responses from authoritative accounts from those that do not. We find that regular users’ popularity and official accounts’ level of busyness do not seem to affect the likelihood of tweets receiving a response.We thus explore the linguistic features of the tweets’ content. Topic modeling and sentiment analysis results suggest that these linguistic aspects of the tweets may affect the response rate from authoritative sources. Our findings suggest crisis-related policy implications, as well as design implications for social media platforms where such exchanges take place, which can potentially increase the reach of credible information in a crisis and help those affected obtain the safety-critical information they need.
https://doi.org/10.1145/3449209
The 24th ACM Conference on Computer-Supported Cooperative Work and Social Computing