Crisis Messaging Journeys: Epistemic Struggles over CDC Guidance During COVID-19

2025-09-13

AI Summary

This study investigates cognitive dissonance, political polarization, and sustained public engagement following CDC’s pandemic-related guidance disseminated via Twitter during COVID-19. Drawing on two years of tweets and reply threads, it introduces the “crisis information journey” framework to examine how archived official statements perpetually reignite public controversy and trust negotiations. Methodologically, the work integrates BERTopic for thematic modeling, VADER for sentiment analysis, Iffy News for credibility assessment, PELT for changepoint detection, and survival analysis for engagement duration modeling—enabling multidimensional, dynamic analysis. Results show that skeptical discourse significantly prolongs user engagement, whereas affirmative responses accelerate disengagement; delayed institutional replies exacerbate polarization. The study innovatively unifies information lifecycle theory, dynamic trust modeling, and behavioral engagement analysis, proposing “responsive communication design” as a practical strategy to strengthen public trust in crisis contexts. (149 words)

Technology Category

• Application Domains: Humanities & Computational Social Science
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization
• Responsible Web: Mitigating misinformation and disinformation
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

This study investigates how the U.S. Centers for Disease Control and Prevention (CDC) communicated COVID-19 guidance on Twitter and how publics responded over two years of the pandemic. Drawing on 275,124 tweets mentioning or addressing @CDCgov, I combine BERTopic modeling, sentiment analysis (VADER), credibility checks (Iffy Index), change point detection (PELT), and survival analysis to trace three phases of discourse: (1) early hoax claims and testing debates, (2) lockdown and mask controversies, and (3) post-vaccine variant concerns. I introduce the concept of crisis messaging journeys to explain how archived "receipts" of prior CDC statements fueled epistemic struggles, political polarization, and sustained engagement. Findings show that skeptical, cognitively complex discourse particularly questioning institutional trust prolonged participation, while positive affirmation predicted faster disengagement. I conclude with design recommendations for annotated, cautious, and flashpoint-responsive communication strategies to bolster public trust and resilience during protracted health crises.

Problem

• Analyzing CDC's COVID-19 Twitter communication and public response patterns
• Examining how archived CDC statements fueled epistemic struggles and polarization
• Investigating how skeptical discourse prolonged public engagement with CDC guidance

Innovation

• BERTopic modeling for discourse analysis
• VADER sentiment analysis application
• PELT change point detection