Content or Context : Which Carries More Weight in Predicting Popularity of Tweets in China

Lun Zhang,Tai-Quan Peng,Ya-Peng Zhang,Xiao-Hong Wang

2012-01-01

Abstract:Through writing short tweets in microblogging sites, millions of users document their life, provide commentary and opinions, express deeply felt emotions, and articulate ideas. How does the content generated by a user compete with the content by others to attract limited human attention? With a framework developed from heuristic-systematic model of information processing, the study aims to uncover content factors and contextual factors that will affect the popularity of tweets. In the study, the popularity of tweets is decomposed into two dimensions: width of distribution and depth of deliberation. The data of the study are 10,000 tweets randomly drawn from a popular microblogging website in China. It is found that content and contextual factors play equally important roles in predicting width of distribution of tweets, while content factors outperform contextual ones in predicting depth of deliberation of tweets. Specifically, topics of tweets and availability of supplementary information are two important content factors in predicting popularity of tweets, while user characteristics (i.e., number of followers and user types) are important contextual factors. Our findings also suggest that re-tweeting and commenting are two distinct behaviors in the context of microblogging.