Framing the Pandemic on Persian Twitter: Gauging Networked Frames by Topic Modeling

Abstract

This study makes a dual contribution to the current literature. First, it examines how Iranian Twitter users framed the COVID-19 crisis in collaborative practice, networked framing. Second, it explores the potential for topic modeling in automated frame identification. The study analyzes a dataset of 4,165,177 tweets collected from Iranian Twittersphere between January 21, 2020 and April 29, 2020. The results indicate that Iranians predominantly framed the pandemic through a political lens and utilized anti-regime networked frames to contest the political system in general and during the pandemic. Furthermore, the study finds that while Latent Dirichlet Allocation (LDA) can accurately identify the most significant networked frames, it may overlook less prominent frames. The research also suggests that LDA performs better with larger datasets and lexical semantics. Lastly, the implications and limitations of the investigation are discussed.