Textual network analysis: Detecting prevailing themes and biases in international news and social media

Abstract

AbstractThe growing volumes of information available today provide opportunities but also challenges for social scientists. This paper presents textual network analysis—a network analysis procedure that transforms any given text into a visual map of words co‐occurring together. It aims at scholars and students who are not familiar with network analysis, showing step‐by‐step how to use this approach and highlighting its advantages and applications. It demonstrates how to identify the main themes appearing in the text as well as to detect its biases and frames. Researchers can use this procedure as a grounded content analysis to formulate theories or as a basis to test existing hypotheses. The second part of the paper presents two studies that applied textual network analysis: (a) to identify the main themes raised by elite newspapers on the “fake news” discourse and (b) to map the topics related to China on Twitter. Both examples show how textual network analysis can be relevant for communication, international relations, and political science scholars as well as for practitioners, wishing to understand the prevailing discourses and tailor their messages more effectively.