Partisan Differences in Legislators' Discussion of Vaccination on Twitter During the COVID-19 Era: Natural Language Processing Analysis

Eden Engel-Rebitzer, Daniel C. Stokes, Zachary F. Meisel, Jonathan Purtle, Rebecca Doyle, Alison M. Buttenheim

Research output: Contribution to journalArticlepeer-review

Abstract

Background: The COVID-19 era has been characterized by the politicization of health-related topics. This is especially concerning given evidence that politicized discussion of vaccination may contribute to vaccine hesitancy. No research, however, has examined the content and politicization of legislator communication with the public about vaccination during the COVID-19 era. Objective: The aim of this study was to examine vaccine-related tweets produced by state and federal legislators during the COVID-19 era to (1) describe the content of vaccine-related tweets; (2) examine the differences in vaccine-related tweet content between Democrats and Republicans; and (3) quantify (and describe trends over time in) partisan differences in vaccine-related communication. Methods: We abstracted all vaccine-related tweets produced by state and federal legislators between February 01, 2020, and December 11, 2020. We used latent Dirichlet allocation to define the tweet topics and used descriptive statistics to describe differences by party in the use of topics and changes in political polarization over time. Results: We included 14,519 tweets generated by 1463 state legislators and 521 federal legislators. Republicans were more likely to use words (eg, "record time,""launched,"and "innovation") and topics (eg, Operation Warp Speed success) that were focused on the successful development of a SARS-CoV-2 vaccine. Democrats used a broader range of words (eg, "anti-vaxxers,""flu,"and "free") and topics (eg, vaccine prioritization, influenza, and antivaxxers) that were more aligned with public health messaging related to the vaccine. Polarization increased over most of the study period. Conclusions: Republican and Democratic legislators used different language in their Twitter conversations about vaccination during the COVID-19 era, leading to increased political polarization of vaccine-related tweets. These communication patterns have the potential to contribute to vaccine hesitancy.

Original languageEnglish (US)
Article numbere32372
JournalJMIR Infodemiology
Volume2
Issue number1
DOIs
StatePublished - Jun 2022

Keywords

  • COVID-19
  • hesitancy
  • linguistic
  • natural language processing
  • NLP
  • partisanship
  • pattern
  • politicization communication
  • Social media
  • Twitter
  • vaccination
  • vaccine

ASJC Scopus subject areas

  • Health Informatics
  • Health Policy
  • Health Information Management
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'Partisan Differences in Legislators' Discussion of Vaccination on Twitter During the COVID-19 Era: Natural Language Processing Analysis'. Together they form a unique fingerprint.

Cite this