Emilio Ferrara and Zeyao Yang have attempted to measure and analyze emotional contagion in Twitter. (Illustration by Suha Park)
With this in mind, Emilio Ferrara, research assistant professor in the Department of Computer Science and computer scientist at the USC Information Sciences Institute, and Zeyao Yang of Indiana University have attempted to measure and analyze emotional contagion in Twitter. For this purpose, they selected a random sample of 3,800 Twitter users, and followed their activity for one week.
“According to the theory of emotional contagion, emotions and emotional states are transferred from one person to another by social interactions,” said Ferrara, who joined USC’s Information Science Institute last year. “We are trying to see if this is also true in the case of social media interactions.”
Ferrara and Yang used an algorithm that computed the ‘emotional’ value of users’ tweets, and classified them as positive, negative, or neutral. They then established comparisons between a user’s tweet and the sentiment of the tweets that appeared in his or her newsfeed during the previous hour.
The study’s results, which were published in the journal PLOS One in November, suggest that positive emotions spread about two times broader on social media than negative ones. This has significant implications in multiple settings such as managing mood disorders.
“What we share on social media matters,” Ferrara said. “We now know that there is a possibility of affecting people’s mood by changing what they are exposed to, and exposing them to more positive messages.”
The research also indicates that a higher number of positive tweets in users’ newsfeeds result in the production of positive messages, while the presence of negative tweets results in the production of negative ones. Furthermore, about 20 percent of Twitter users were classified as highly susceptible to emotional contagion — with more than half of their tweets affected. Those users were four times more likely to be affected by positive tweets than negative ones.
“Similarly to what happens with people’ susceptibility to catching the flu, some individuals are more inclined to change emotional states as a reaction to the contents they are exposed to, while others hold stronger emotional states that rarely change irrespectively of the type of emotional stimuli they receive,” Ferrara said.
The study’s results could help determine ways of better communicating messages in social media, and given its undeniable impact in modern life, insights on online social interactions are most welcome in every field.
“We can already see encouraging signals in people’s relationship with information found online,” Ferrara said. “For example, more and more people are now able to realize when they are exposed to rumors or inaccurate reporting online.”