Delineating the Transnational Network Agenda-Setting Model of Mainstream Newspapers and Twitter: A Machine-Learning Approach

Abstract

ABSTRACT The months-long anti-extradition bill movement in Hong Kong has gained worldwide attention. Grounded in the network agenda-setting (NAS) model, this study utilizes a machine-learning approach to analyze the related coverage of mainstream newspapers in Hong Kong, Mainland China, the U.S. and the U.K. (N = 2118), as well as discussions on Twitter (N = 152,509). Network visualizations showed that each media utilized a unique approach to highlight and connect the substantive and affective attributes. Time-series analyses revealed an overall reciprocal whilst asymmetrical association between the newspapers and Twitter, in which the latter exhibited a stronger influence on the former, particularly in terms of substantive attribute agendas. Yet, Twitter’s impact shrank in terms of the affective attribute agendas and the NAS models. Newspapers, though exerted rather limited impact on Twitter, maintained a certain extent of independence in setting their affective attribute agendas and NAS models. This study enriches the NAS literature through combining substantive and affective attributes in semantic networks.