In Search of Public Agenda with Text Mining: An Exploratory Study of Agenda Setting Dynamics Between the Traditional Media and Wikipedia

Abstract

When Downs [1] proposed his famous issue-attention cycle in 1972, he thought that the mass media would report news and information that arouses people’s interests. This thought, however, is prone to challenges. With the prevalence of the Internet and, perhaps more importantly, the concept of Web 2.0, Wikipedia becomes another major source of information for the public. Given that Wikipedia allows anyone to edit the content, the details about a particular event or issue on pages of Wikipedia can be considered as a quasi-public agenda. Understanding this quasi-public agenda may help us evaluate different models of policy cycles, including the Downs’ famous issue-attention cycle. My study aims to assess the agenda setting dynamics among 5 major news outlets in the UK, as the traditional mass media, and Wikipedia, as a form of participatory journalism. By agenda, it refers to the choices of frames and sentiment. Using text mining techniques, my study assesses the choices of frames and sentiment adopted by the articles of the news outlets and the Wikipedia pages concerning with the issue Brexit. The timeline of the study is between the date when the Wikipedia page “Brexit” emerged and the date of the Brexit referendum. The study also explores the possible relationship between these agendas. Frame analysis of the news articles will be conducted through automatic text classification, whereas the frames on the Wikipedia pages will be analyzed through both text classification and clustering. Lexicon-based approach will be used for sentiment analysis. The relationship between the agendas will be explored through Granger-causality tests.