This is a small RAPID award to provide support for a project involving the analysis of the emotional content of large-scale Twitter data to shed light on important and enduring theoretical questions in Social Psychology. The investigators have expertise in Social Informatics and Social Psychology, and together they will collect and analyze a large-scale collection of tweets (short messages broadcast using the Twitter social networking service) that reference the Haiti earthquake disaster. The emotional content of these tweets will be tracked over time in an automated fashion using an expanded version of the Profile of Mood States test previously developed by one of the PIs. The social networks connecting Twitter users will be traced from users' "Friend of" and "Follower" data which will enable the researchers to examine the bi-directional influence of social factors with emotional responses and prosocial behavior. This analysis of Twitter data (both tweet content and social networking data of Follower and Friends networks) will permit an analysis of the impact of emotions on donating or other forms of prosocial behavior, as well as the effect of donating on subsequent emotional states. Using Twitter data to address what represents longstanding theoretical issues in Social Psychology allows the researchers to avoid many of the limitations of laboratory-based studies that include generally short time duration, practical and ethical limitations on the use of high impact manipulations, and often a reliance on college undergraduates as participants.
The broader impacts of this work are two-fold. First, the proposed program of research will promote teaching, training, and learning, by training and mentoring undergraduate research assistants and graduate students. Second, the proposed research will address questions that can be applied more broadly, to understand why and how people donate or perform other prosocial behaviors, and to encourage them to do so more frequently. Knowing what emotional states tend to lead to donation, and how people respond to information about their friends or acquaintances donations will allow charitable organizations to construct more effective appeals.
Soon after the news of the Haiti earthquake disaster started to spread, people used the Twitter network to share their feelings about the disaster and organize to aid the afflicted. In fact, a good deal of the charitable response to the Haiti disaster took place on Twitter. Aid organizations and individuals leveraged the "viral" nature of its infrastructure to maximize awareness and issue appeals for donations and assistance. In this research project we used computers to analyse the emotional content of millions of Tweets and the social connections between users to determine (1) whether the emotions of Twitter users are affected by the emotions of other users, and (2) whether their intentions to donate are influenced by the intentions that others mention in their Tweets. We found that users who are connected on Twitter are indeed more likely to share the same emotions [1], a phenomenon referred to as "assortative mixing" in social network research. Preliminary results also suggest that online social influence indeed affects the tendency of Twitters to express the intent to engage in charitable behavior. Beyond addressing social influence over emotions and helping behavior, both long-standing interests within Social Psychology, the results of this research project have numerous societal applications. The way in which news, emotions, and various other communications spread from one individual to the next in online social networks has often been described as "viral" and could lead to unpredictable outcomes. A better understanding of how factors such as social influence and emotions play a role in our collective response to large-scale disasters may contribute to help organizations, and society as a whole, better organize their online communications and coordinate efforts in the case of emergencies and disasters. [1] Johan Bollen, Bruno Gonçalves, Guangchen Ruan and Huina Mao. Happiness Is Assortative in Online Social Networks. Artificial Life, Summer 2011, Vol. 17, No. 3, Pages 237-251
