How COVID-19 has Impacted the Anti-Vaccine Discourse: A Large-Scale Twitter Study Spanning Pre-COVID and Post-COVID Era

TL;DR

The paper develops two label-semantics–aware classifiers, CoV-Ent and CoV-Gen, to detect fine-grained anti-vaccine concerns on Twitter and applies them to a 5-year tweet corpus (2018–2023) to track how COVID-19 reshaped anti-vax discourse. By leveraging the CAVES dataset with 11 specific concerns plus a None label, the authors show that COVID-19 broadened and intensified the anti-vaccination argument space, and that COVID-related concerns increasingly spill over to non-COVID vaccines. They also identify traditional and converted anti-vaxxers and analyze how their concerns evolve post-pandemic, revealing persistent and new anxieties about non-COVID vaccines such as the Flu vaccine. The work offers practical tools for real-time, personalized counter-messaging and highlights methodological avenues for large-scale, semantics-aware analysis of social-media misinformation in public health contexts.

Abstract

The debate around vaccines has been going on for decades, but the COVID-19 pandemic showed how crucial it is to understand and mitigate anti-vaccine sentiments. While the pandemic may be over, it is still important to understand how the pandemic affected the anti-vaccine discourse, and whether the arguments against non-COVID vaccines (e.g., Flu, MMR, IPV, HPV vaccines) have also changed due to the pandemic. This study attempts to answer these questions through a large-scale study of anti-vaccine posts on Twitter. Almost all prior works that utilized social media to understand anti-vaccine opinions considered only the three broad stances of Anti-Vax, Pro-Vax, and Neutral. There has not been any effort to identify the specific reasons/concerns behind the anti-vax sentiments (e.g., side-effects, conspiracy theories, political reasons) on social media at scale. In this work, we propose two novel methods for classifying tweets into 11 different anti-vax concerns -- a discriminative approach (entailment-based) and a generative approach (based on instruction tuning of LLMs) -- which outperform several strong baselines. We then apply this classifier on anti-vaccine tweets posted over a 5-year period (Jan 2018 - Jan 2023) to understand how the COVID-19 pandemic has impacted the anti-vaccine concerns among the masses. We find that the pandemic has made the anti-vaccine discourse far more complex than in the pre-COVID times, and increased the variety of concerns being voiced. Alarmingly, we find that concerns about COVID vaccines are now being projected onto the non-COVID vaccines, thus making more people hesitant in taking vaccines in the post-COVID era.

Figures (4)

• Figure 1: Distribution of concerns in anti-vax tweets across the different time-periods. The 'side-effect' class (the largest class) is shown separately, to enable better visualization of the other classes. The inset graph shows the decline of the ineffective class in post-COVID period.
• Figure 2: Distribution of concerns among anti-vax tweets about non-covid vaccines across the different time-periods.
• Figure 3: Distribution of concerns in tweets from post-COVID, (i) that mention both COVID and non-COVID vaccines, and (ii) that mention only non-COVID vaccines.
• Figure 4: Distribution of anti-vaccine concerns posted by the traditional anti-vaxxers during pre- and post-COVID periods (red and blue bars), and the converted anti-vaxxers during post-COVID (yellow bars).