Role of Sentiment in Message Propagation: Reply vs. Retweet Behavior in Political Communication

This paper examines the role of sentiment in information propagation. We make use of political communication in the Twitter space, and relate emotion expressions in a message to the degrees of responses generated by the message. We also compare differences between user reply vs. retweet behavior with respect to sentiment variables. The current results indicate that that degree of emotion expressions in twitter messages can affect the number of replies generated as well as retweet rates. Due to the difference in the nature of endorsement (retweet) vs. responses (replies or conversation), some of the variables present opposite roles in explaining the degree of responses the message receives. We expect these results will help generating a predictive model of message propagation.