Previous research suggests the existence of sentiments in online social networks. In comparison to real life human interaction, in which sentiments have been shown to have an influence on human behaviour, it is not yet completely understood which mechanisms explain how sentiments influence users in online environments. We develop a theoretical framework that tries to bridge the gap between social influence theories that focus on offline interactions on one hand and online interaction in social networks on the other hand. We then test our hypothesis about the influence and dissemination of sentiments in a quantitative analysis that is based on retrieved textual messages of communication patterns in over 12000 online social networks. Our empirical results suggest a general influence of sentiments on node communication patterns that is evidenced by increased occurrences of subsequent messages that express the same sentiment polarization. We interpret these findings and suggest future research to advance our currently limited theories that assume perceived and generalized social influence to path-dependent social influence models that consider actual behaviour.
Ecis 2013 Proceedings, 2013
Social Network Analysis; Sentiment Analysis; Communication Patterns
Main Research Area:
Proceedings / European Conference on Information Systems (ecis)
The 21st European Conference on Information Systems (ECIS) 2013
Association for Information Systems. AIS Electronic Library (AISeL)