Politicization of Science in COVID-19 Vaccine Communication: Comparing US Politicians, Medical Experts, and Government Agencies

ABSTRACT

We compare the social media discourses on COVID-19 vaccines constructed by U.S. politicians, medical experts, and government agencies, and investigate how various contextual factors influence the likelihood of government agencies politicizing the issue. Taking the political corpus and the medical corpus as two extremes, we propose a language-based definition of politicization of science and measure it on a continuous scale. By building a machine learning classifier that captures subtle linguistic indicators of politicization and applying it to two years of government agencies’ Facebook posting history, we demonstrate that: 1) U.S. politicians heavily politicized COVID-19 vaccines, medical experts conveyed minimal politicization, and government agencies’ discourse was a mix of the two, yet more closely resembled medical experts;’ 2) increasing COVID-19 infection rates reduced government agencies’ politicization tendencies; 3) government agencies in Democratic-leaning states were more likely to politicize COVID-19 vaccines than those in Republican-leaning states; and 4) the degree of politicization did not significantly differ across agencies’ jurisdiction levels. We discuss the conceptualization of politicization of science, the incumbency effect, and government communication as an emerging area for political communication research.

KEYWORDS:

• Politicization of science
• COVID-19
• social media
• government communication
• computational methods

Acknowledgments

We are grateful to editors, anonymous reviewers, Kaiping Chen, Kokil Jaidka, Yphtach Lelkes, Erik Nisbet, Michael Xenos, Tian Yang, as well as seminar participants at MPSA and the University of Pennsylvania’s Democracy & Information Group (DIG). All remaining errors are our own. Replication materials associated with the analyses have been deposited onto Open Science Framework at https://doi.org/10.17605/OSF.IO/RMBGKhttps://doi.org/10.17605/OSF.IO/RMBGK

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/10584609.2023.2201184

Notes

1. We opted for logistic regressions over other methods to provide a more accessible model for the wider social scientific and public communities (Theocharis & Jungherr, 2021). The classifier puts different weights (as coefficients) on different words (tokens), constitutes a transparent and interpretable algorithm, and provides more granularity than the keyword-matching method employed by prior studies. Other methods, such as SVM and BERT, produce more black-box models that are hard to interpret (Molnar, 2019). Our goal is not to compare the accuracy rates of different classifiers, but to demonstrate the utility of measuring politicization of science from a linguistic perspective on a continuous scale from 0 to 1, for which logistic regressions are the best fit.

2. https://www.ncsl.org/research/about-state-legislatures/partisan-composition.aspx.

3. https://github.com/nytimes/covid-19-data Few states changed their reporting protocols during the COVID-19 pandemic, which led the state’s cumulative number of positive cases to reduce in some weeks. We excluded observations with negative numbers of weekly new cases from our regression models.

4. We provided this classifier in our replication materials, however, it should be noted that the classifier was trained for the specific context of COVID-19 vaccine communication. Researchers are advised to adjust our script to re-train classifiers for their own studies.

Additional information

Notes on contributors

Alvin Zhou

Alvin Zhou (Ph.D., University of Pennsylvania) is an Assistant Professor at the Hubbard School of Journalism and Mass Communication at the University of Minnesota. His research centers around computational social science and strategic communication (organizations, advertising, and public relations).

Wenlin Liu

Wenlin Liu (Ph.D., University of Southern California) is an Assistant Professor in the Jack J.Valenti School of Communication at the University of Houston. Her research focuses on social media-mediated disaster communication, multiethic community building, and social networks.

Aimei Yang

Aimei Yang (Ph.D., University of Oklahoma) is an associate professor of public relations in the Annenberg School for Communication and Journalism at the University of Southern California. Yang’s research is positioned at the intersection of strategic advocacy, social-mediated networks, and civil society research.