Micro-blogging website Twitter can be utilised to gauge negative emotions in the aftermath of violence, especially on campuses since students are the most likely users among the general population, revealed a research.
Twitter can help analysts to gather the overall effect of traumatic experiences, like a school shooting, on people.
“Twitter’s rapid distribution and widespread use enable us to avoid the fundamental difficulties with traditional survey methodologies,” said lead researcher Roxane Cohen Silver, Professor at the University of California, Irvine.
“Studying communities impacted by traumatic events is often costly and requires swift action to enter the field when disaster strikes,” Silver added.
Social media, such as Twitter, offers a wealth of information that can provide an immediate window into a community’s emotional response to a trauma experienced collectively by them. However, using a big data approach presents its own unique challenges, according to Nickolas M. Jones, student at University of California, Irvine.
“Locating community members who have experienced the trauma can be problematic,” Jones explained.
To infer members’ location, prior researchers have either used geotagged tweets, which account for only 3 to 6 per cent of Twitter users, or tweets with hashtags, which yields vast numbers of tweets without certainty of the users’ location.
The new approach was deployed to study the impact of the deadly gun violence at University of California-Santa Barbara, Northern Arizona University and Oregon’s Umpqua Community College.
Followers of local Twitter accounts were identified, and the tweets of these likely community members were then downloaded for data analysis.
In the study, the team compared negative emotional expression via Twitter by people in the impacted communities and those in three matched control communities shortly before and after the campus killings.
Despite variations in the severity of violence at the colleges, results showed similar patterns of increased negative emotional expression in the post-event tweets from those in the affected areas, while tweets from those in the matched control locales exhibited no change in negative emotional expression during the same period of time.
The study features in the forthcoming issue of the journal Psychological Methods.