COVID-19 Misinformation and Infodemic¶
During the COVID-19 pandemic, false and misleading information about the virus, vaccines, treatments, and public health measures proliferated across social media, messaging apps, and traditional media. This rapid spread of problematic information—the infodemic—has been a major research focus due to its documented harms on public health behavior and policy.
Key dimensions of COVID-19 misinformation research:
Platform dynamics — Information spreads differently across social media platforms, with some amplifying unreliable sources more than others. Early pandemic data shows R₀ values (basic reproduction numbers for information) exceeding epidemic thresholds across Twitter, Instagram, YouTube, Reddit, and Gab.
Vaccine hesitancy — Vaccine-related misinformation has been a primary focus, studying susceptibility factors, intervention effectiveness, and the role of social networks in vaccine confidence.
Institutional responses — Fact-checking, content moderation, and official health communication have played roles in countering misinformation spread.
Temporal dynamics — Information diffusion patterns shifted as pandemic evolved, with different topics dominating at different times.
Key papers¶
- A large-scale COVID-19 Twitter chatter dataset for open scientific research - an international collaboration — Large-scale curated corpus of 800+ million COVID-19-related tweets (Jan–Nov 2020) with multilingual coverage and FAIR principles compliance; designed to support misinformation identification, sentiment analysis, and cascade dynamics research; maintained with biweekly updates and comprehensive preprocessing pipelines
- Memon & Carley (2020) — Characterizing COVID-19 Misinformation Communities — Twitter dataset (CMU-MisCOV19) with 4,573 annotated tweets across 17 categories; characterizes misinformed vs informed communities by network density, bot prevalence (19% vs 11%), sociolinguistic patterns, and vaccination stance; shows 41% of misinformed users are anti-vaxxers
- Fighting an Infodemic: COVID-19 Fake News Dataset — Dataset of 10,700 COVID-19 posts/articles with binary labels; benchmarks ML baselines achieving 93.32% F1-score with SVM
- Cinelli et al. (2020) — The COVID-19 Social Media Infodemic — Multi-platform analysis of information spreading with epidemic models; shows platform-dependent amplification of misinformation
- Pennycook et al. (2020) — Fighting COVID-19 misinformation on social media — Testing accuracy-focused interventions during pandemic
- Roozenbeek et al. (2020) — Susceptibility to misinformation about COVID-19 — Individual and situational risk factors
- Guess et al. (2020) — A digital media literacy intervention — Intervention testing with health and political misinformation
- A Heuristic-driven Uncertainty based Ensemble Framework for Fake News Detection in Tweets and News Articles — Ensemble of pre-trained language models for COVID-19 fake news detection; achieves F1=0.9892 on CONSTRAINT COVID-19 Fake News dataset using transfer learning, soft voting, statistical feature fusion, and uncertainty quantification
Related topics¶
- Infodemic — broader phenomenon of information excess and disorder
- Misinformation — false information more generally
- Health misinformation — medical and health-related false claims
- Vaccine hesitancy and vaccine safety concerns — specific to vaccine-related misinformation