Politics Go “Viral”: A Computational Text Analysis of the Public Attribution and Attitude Regarding the COVID-19 Crisis and Governmental Responses on Twitter

Abstract

The U.S. confronts an unprecedented public health crisis, the COVID-19 pandemic, in the presidential election year in 2020. In such a compound situation, a real-time dynamic examination of how the general public ascribe the crisis responsibilities taking account to their political ideologies is helpful for developing effective strategies to manage the crisis and diminish hostility toward particular groups caused by polarization. Social media, such as Twitter, provide platforms for the public’s COVID-related discourse to form, accumulate, and visibly present. Meanwhile, those features also make social media a window to monitor the public responses in real-time. This research conducted a computational text analysis of 2,918,376 tweets sent by 829,686 different U.S. users regarding COVID-19 from January 24 to May 25, 2020. Results indicate that the public’s crisis attribution and attitude toward governmental crisis responses are driven by their political identities. One crisis factor identified by this study (i.e., threat level) also affects the public’s attribution and attitude polarization. Additionally, we note that pandemic fatigue was identified in our findings as early as in March 2020. This study has theoretical, practical, and methodological implications informing further health communication in a heated political environment.

Mobilization, self-expression or argument? A computational method for identifying language styles in political discussion on Twitter

“…Li et al, 2023), attributions (e.g. Zhang et al, 2023) and intentions (e.g. Xu et al, 2023), it necessitates pre-labeled training data, which can be challenging to acquire (Hu et al, 2022).…”

“…However, while Twitter's democratizing tendencies can exert a positive impact on certain political occurrences, a growing body of research also underscores its role in accentuating intergroup disparities (Hu and Kearney, 2021), fostering prejudice (Postmes et al, 2002;Carr, 2018) and perpetuating polarization (Zhang et al, 2023) due to the deindividuation effect (Postmes et al, 2002) and the selective exposure (Spohr, 2017) inherent in computer-mediated communication. Consequently, Twitter serves as a virtual arena wherein the demarcation between distinct political factions often becomes more conspicuous and intense than in offline interactions.…”

Biological changes, political ideology, and scientific communication shape human perceptions of pollen seasons

“…Social media has been shown to provide powerful big data for public perceptions of climate change (Jang & Hart, 2015) . Further, the rich textual information gained from social media can provide in-depth insights into public discourses related to climate change (Kim et al, 2021;W. Zhang et al, 2023;Ghermandi et al, 2023;Anderson, 2017) .…”

A Two-Step Method for Classifying Political Partisanship Using Deep Learning Models

“…Political partisanship is one of the most predominant group identities in society. It has a powerful impact on various political attitudes and behaviors within both developed and developing democracies (Van Bavel & Pereira, 2018) and can be used to explain political engagement, partisan reasoning, voting patterns (Huddy & Bankert, 2017; Huddy et al, 2015), and communication patterns (Xu et al, 2023; Zhang et al, 2023). People within the same political affiliation usually have shared patterns of behaviors, thoughts, feelings, and expressions.…”

“…As a result of affiliating with different groups, individuals tend to have different political, ideological, and cultural worldviews to see and interpret the world (Carney et al, 2008; Shook & Fazio, 2009). These differences can be identified in discussions regarding public issues such as abortion (e.g., Carmines et al, 2010), gun control (e.g., Pearson-Merkowitz & Dyck, 2017), climate change (e.g., Dunlap et al, 2016), and even the COVID-19 pandemic (Zhang et al, 2023).…”