Global misinformation spillovers in the online vaccination debate before and during COVID-19

Jacopo Lenti, Kyriaki Kalimeri, André Panisson, Daniela Paolotti, Michele Tizzani, Yelena Mejova, Michele StarniniarXiv 2022 · DOI · arXiv

Introduction

The COVID-19 pandemic transformed vaccine hesitancy from a localized concern into a global "infodemic." While previous research often focused on single-country or English-only datasets, this study provides a large-scale, multi-lingual analysis of how vaccine misinformation travels across national borders. By analyzing over 300 million tweets, the researchers demonstrate that anti-vaccination communities became increasingly interconnected globally, with the United States serving as a primary exporter of low-credibility information.

Methodology

The researchers employed a "human-in-the-loop" network science approach to map the discourse:

  • Network Reconstruction: Two types of networks were built for 28 countries: Retweet (RT) networks to capture social interaction and Co-sharing (CO) networks to identify shared information sources.
  • Hierarchical Clustering: The Paris algorithm was used to identify large, stable communities within the data, avoiding the fragmentation common in standard unsupervised algorithms.
  • Polarization Metrics: The Random Walk Controversy (RWC) score measured the degree of "siloing" in the debate. High RWC indicates that users are unlikely to encounter opposing viewpoints.
  • Echo Chamber Analysis: Normalized Mutual Information (NMI) compared RT and CO networks. High NMI suggests that social groups are consuming identical information sources, a key indicator of an echo chamber.

Data Overview

The study utilized a massive dataset consisting of 316 million vaccine-related tweets in 18 languages, collected between October 2019 and March 2021. Approximately 48.7% of users were geolocated using the GeoNames database. To track misinformation, the authors compiled a list of 1,732 "low-credible" domains identified by fact-checking organizations in Italian, French, English, and Greek.

Key Results

Temporal Evolution and Surge in Volume

The volume of the vaccination debate increased by two orders of magnitude as the pandemic progressed. The timeline reveals distinct phases, with major spikes in activity corresponding to key vaccine milestones such as the announcement of the Sputnik V, Pfizer, and Moderna vaccines.

Volume of the vaccination debate on Twitter across different phases of the pandemic. Spikes correspond to major events like the announcement of the Sputnik V (A), Pfizer (B), and Moderna (C) vaccines.

The Global No-Vax Network

Anti-vaccination views are not isolated; they form a "tightly-knit global no-vax network." These communities are more likely to share URLs and YouTube videos than the general population. The study found that the U.S. is the central "net exporter" of vaccine misinformation, with 68% of all low-credible URLs retweeted worldwide originating from U.S.-based accounts.

The "Jan 6th" Effect: Intersection of Politics and Health

A significant finding is the impact of political content moderation on health discourse. Following the January 6th U.S. Capitol attack, the suspension of accounts associated with political extremism led to a massive, worldwide drop in vaccine misinformation. The proportion of low-credible domains in URLs retweeted from the U.S. fell from 14% to 7%. This suggests that the same "super-spreaders" often drive both political and health-related misinformation.

International Dynamics

During the vaccine rollout, Russian users emerged as strategic exporters of information, particularly targeting Latin American countries with campaigns like #sputnikesesperanza. Overall, the no-vax communities showed high suspension rates (13.3%), nearly 7.5 times higher than the rest of the network.

Limitations and Future Work

While comprehensive, the study has limitations. Twitter users are not perfectly representative of the global population, and the availability of "low-credible" domain lists was limited to four languages, likely making the findings a lower bound of the actual misinformation flow. Future research should aim to include more data from Africa and Asia to complete the global map of health misinformation.

Summary updated

Abstract

Anti-vaccination views pervade online social media, fueling distrust in scientific expertise and increasing vaccine-hesitant individuals. While previous studies focused on specific countries, the COVID-19 pandemic brought the vaccination discourse worldwide, underpinning the need to tackle low-credible information flows on a global scale to design effective countermeasures. Here, we leverage 316 million vaccine-related Twitter messages in 18 languages, from October 2019 to March 2021, to quantify misinformation flows between users exposed to anti-vaccination (no-vax) content. We find that, during the pandemic, no-vax communities became more central in the country-specific debates and their cross-border connections strengthened, revealing a global Twitter anti-vaccination network. U.S. users are central in this network, while Russian users also become net exporters of misinformation during vaccination roll-out. Interestingly, we find that Twitter's content moderation efforts, and in particular the suspension of users following the January 6th U.S. Capitol attack, had a worldwide impact in reducing misinformation spread about vaccines. These findings may help public health institutions and social media platforms to mitigate the spread of health-related, low-credible information by revealing vulnerable online communities.
BibTeX
@misc{lenti2022global,
  author = {Jacopo Lenti and Kyriaki Kalimeri and André Panisson and Daniela Paolotti and Michele Tizzani and Yelena Mejova and Michele Starnini},
  title = {{Global misinformation spillovers in the online vaccination debate before and during COVID-19}},
  year = {2022},
  eprint = {2211.11495},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2211.11495},
}
Figure