Predicting police micro-blog post diffusion on Twitter™: the man or the message?

Abstract

Message diffusion on micro-blogs such as Twitter™ is still poorly understood. This study aims to develop an explanatory model of message diffusion to understand why some messages are forwarded and others are not. By studying 934 Dutch police force accounts, this study tests previous insights using trace data drawn from the Twitter™ API. Based on an iterative human-calibration procedure, message topics were automatically coded based on customized lexicons. A principal component analysis of message characteristics generated four distinct patterns of use. Message characteristics where combined with user characteristics in a multilevel logistic general linear model. Main results show that URL or use of informal communication increase chances of message forwarding. In addition, contextual factors such as user characteristics impact diffusion probability. Recommendations are discussed for further research into authorship styles and their implications for message diffusion. For practitioners, a list of recommendation about how to increase message reach is presented.