A computational analysis of agenda setting and second-level agenda setting

Abstract

Agenda setting theory explains how media affects its audience. Limitations of traditional media studies about agenda setting include the small set of issues, the costly survey data of public interest, and the expertise needed for categorizing the article frames. In this paper, I propose a computational approach to study agenda setting with a large dataset of online news, approximately 17,000 articles from the website of National Public Radio. Along with the articles, I crawl user comments and social sharing counts. With that data, I automatically extract the major issues with hierarchical Dirichlet processes, a nonparametric probabilistic topic model. Then, I quantify the effects of agenda setting by analyzing the correlation of the user comments and social sharing with the amount of news coverage. Finally, I use sentiment analysis to define and analyze the effects of second-level agenda setting. By using advanced machine learning tools, I demonstrate the potential of detailed and principled analysis of agenda setting from a large set of publicly available data.