Distress and rumor exposure on social media during a campus lockdown

Proc Natl Acad Sci U S A. 2017 Oct 31;114(44):11663-11668. doi: 10.1073/pnas.1708518114. Epub 2017 Oct 17.


During crisis events, people often seek out event-related information to stay informed of what is happening. However, when information from official channels is lacking or disseminated irregularly, people may be at risk for exposure to rumors that fill the information void. We studied information-seeking during a university lockdown following an active-shooter event. In study 1, students in the lockdown (n = 3,890) completed anonymous surveys 1 week later. Those who indicated receiving conflicting information about the lockdown reported greater acute stress [standardized regression coefficient (b) = 0.07; SE = 0.01; 95% confidence interval (CI), 0.04, 0.10]. Additionally, those who reported direct contact with close others via text message (or phone) and used Twitter for critical updates during the lockdown were exposed to more conflicting information. Higher acute stress was reported by heavy social media users who trusted social media for critical updates (b = 0.06, SE = 0.01; 95% CI, 0.03, 0.10). In study 2, we employed a big data approach to explore the time course of rumor transmission across 5 hours surrounding the lockdown within a subset of the university's Twitter followers. We also examined the patterning of distress in the hours during the lockdown as rumors about what was happening (e.g., presence of multiple shooters) spread among Twitter users. During periods without updates from official channels, rumors and distress increased. Results highlight the importance of releasing substantive updates at regular intervals during a crisis event and monitoring social media for rumors to mitigate rumor exposure and distress.

Keywords: Twitter; crisis event; lockdown; rumor; social media.

MeSH terms

  • Emergencies*
  • Humans
  • Information Dissemination*
  • Social Media*
  • Stress, Psychological*
  • Universities*
  • Young Adult