In two new studies, iSchool researchers explore social media’s role in expressions of grief and responses to tragedies.
by Kimberly Glasgow and Jessica Vitak
“Thank you.” What can such simple words help us understand in the face of disaster? Looking at disaster-stricken communities on social media, we can identify expressions of gratitude, or thanks, for social support these communities received to help them cope. Gratitude for support can be thought of not only as an indirect proxy for the social support a community received, but as an indicator of resilience and potential for recovery. The health and resilience of our communities after traumatic events is not easy to measure and monitor, but social media, such as Twitter, could help provide insight.
In two recently published articles, we partnered with HCIL faculty Yla Tausczik and Johns Hopkins University Applied Physics Laboratory researcher Clay Fink (a colleague of Kim’s) to examine how people used Twitter in the aftermath of two localized tragedies. Kim and Clay collected large datasets of tweets following the Sandy Hook Elementary School (Newtown, CT) shooting in December 2012 and an April 2011 tornado in Alabama that ranks as the costliest tornado in U.S. history. We examined nearly two weeks’ worth of tweets following the two events, using machine learning to detect expressions of gratitude for social support. When comparing the two tragedies, we found that expressions of gratitude were significantly higher in Newtown than in Alabama, even though the scope of damage was much higher for those affected by the tornado. Such differences in tweeting patterns may reflect larger, global outpourings of emotional support after what has been characterized as a senseless, human-wrought tragedy—compared with natural weather events where we expect to lack control. For both communities, we found that social media may serve to bridge the distance between the affected community and those outside who are willing to offer support, although offers of support are likely to be higher when there is a stronger connection between those outside the affected community and the event. This paper recently won the Best Conference Paper Award at the 9th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction (SBP-BRiMS) and can be viewed online here.
For our second study, we delved more deeply into the types of support reflected within (tweeted) expressions of gratitude by members of the Newtown community. Through a content analysis of nearly 1200 tweets including expressions of gratitude in the four weeks following the mass shooting, we developed a taxonomy including six types of social support enabled through Twitter’s affordances. Some were familiar and reflected the long history of empirical work on social support. For example, offers of tangible support (e.g., food and other donations) was common. Nurturant support, or providing emotional comfort, was also prevalent. Yet equally meaningful to the community was another type of support rarely discussed in prior literature, which we have termed symbolic support. Symbolic support is expressed in seemingly small acts such as wearing an armband or holding a moment of silence to honor and respect the victims. As one local resident noted, “40 other states observed #momentofsilence thank you. Words still fail us, silence says more. #newtown #sandyhook.” This paper recently won the Best Conference Paper Award at the 2016 International Conference on Social Media & Society (SM+S) and can be viewed online here.
When considering the findings from these two studies, we believe that analyzing organic expressions and responses to events by community members may advance our understanding of the complex linkages between the experience of disaster, support, gratitude, and post-traumatic growth. These findings offer ways for social media – as a window into real-time community behaviors relating to response and healing after disaster – to contribute to the provision of mental health resources and monitoring community resilience and recovery.