A recent study led by researchers from the University of Tokyo has shed light on the intricate web of factors contributing to vaccine hesitancy and resistance. The study, published in the Journal of Computational Social Science, identifies online political engagement, conspiracy theories, and spirituality as significant influences shaping anti-vaccine beliefs across different societal groups. The findings are based on an extensive analysis of nearly 100,000,000 posts on the social media platform Twitter (now known as X).
To conduct their study, Professor Fujio Toriumi from the Department of Systems Innovation and his colleagues first set about identifying Twitter accounts that actively spread anti-vaccine information. This identification process was based on the content of tweets related to vaccines, specifically looking for narratives that were either critical of vaccines or outright opposed to their use.
The team employed the Twitter Public API to gather tweets containing references to vaccines, yielding a substantial dataset spanning the entirety of 2021. This initial collection phase resulted in 98,805,971 tweets, from which the researchers then extracted relevant data for their analysis.
Following the data collection, the researchers embarked on a meticulous process of clustering these tweets to categorize them based on the nature of their content—ranging from pro-vaccine to anti-vaccine sentiments. This clustering involved calculating similarity coefficients between tweets and employing the Louvain method for community detection, a technique well-regarded for its efficacy in identifying natural groupings within large datasets.
Through their methodical clustering of tweets, the researchers found that the conversations around vaccines were not monolithic but rather spanned a spectrum from strong support to vehement opposition. The largest clusters represented these varied perspectives, with the most significant portion of the discourse being pro-vaccine, followed by clusters expressing skepticism about government vaccine policies, and finally, a considerable segment of tweets that were explicitly anti-vaccine.
With these clusters defined, the next step focused on the individuals behind the tweets — specifically, accounts that disproportionately followed those identified as spreading anti-vaccine information.
The analysis of Twitter data revealed that individuals who adopted anti-vaccine beliefs during the pandemic were more likely to follow and engage with accounts promoting conspiracy theories, spiritual narratives, and alternative health practices, illustrating the multifaceted pathways through which these beliefs are formed and reinforced.
These users were also more politically engaged than their counterparts who did not express such strong views on vaccines. Interestingly, the analysis revealed that anti-vaccine sentiments were not confined to a particular political ideology but were prevalent across the spectrum, with a notable presence among both rightist and leftist groups. However, left-leaning anti-vaccine users were more dominant.
“The uniqueness of conspiracy theorists leaning left in Japan may be attributed to the impact of the 2011 Fukushima nuclear disaster,” said Toriumi. “While anti-nuclear sentiments have long been associated with leftist ideologies, the addition of the fear stemming from radioactive contamination has led to the proliferation of conspiracy theories predominantly within the leftist spectrum, believed to be related to fear surrounding the incorporation of foreign substances into the human body. This might have heightened fear, hesitancy and mistrust in vaccines during the pandemic and was likely reinforced by increased representation online.”
The researchers also conducted a longitudinal analysis to understand of how anti-vaccine attitudes evolved over time. By examining the change in the proportion of anti-vaccine information within users’ Twitter feeds from January 2020 to December 2021, the researchers were able to track shifts in sentiment.
A notable increase in the proportion of accounts classified within the high anti-vaccine group was observed, indicating that the pandemic period saw a significant rise in vaccine hesitancy and resistance among Twitter users in Japan. This shift suggests that the unprecedented context of the pandemic, coupled with the pervasive influence of social media, played a critical role in amplifying anti-vaccine narratives.
“During the COVID-19 pandemic, there was a rise in anti-vaccine sentiments on social media, and our study aimed to understand the triggers that led individuals to adopt anti-vaccine attitudes,” said Toriumi. “We found anti-vaccination conspiracy theorists, so-called anti-vaxxers, exhibited stronger political engagement compared to vaccine supporters. Although some Japanese users express right-wing tendencies, a majority lean toward more left-wing ideologies, in contrast to what was observed in the West.”
While the study does not definitively prove causality, it underscores the possibility that conspiracy theories and spirituality may serve as pathways for individuals to gravitate towards more polarizing political figures and parties.
“Spirituality, naturalism, alternative health practices and anti-vaccine sentiments all have something in common: their indifference or even disdain for scientific evidence,” said Toriumi. “Individuals interested in these topics tend to pick what scientific facts suit their opinions. Also, they exhibit strong resistance to the incorporation of artificial substances into their bodies under the guise of naturalism. It is believed that these similarities serve as gateways to anti-vaccine conspiracy theories.”
Another significant aspect of the study was its focus on the emergence of the Sanseito party in Japan, which capitalized on anti-vaccine narratives alongside other issues like anti-immigration rhetoric and environmentalism. The researchers discovered that Sanseito’s ability to gain political traction was significantly bolstered by its appeal to those who harbored anti-vaccine beliefs, especially those who had adopted such views amid the pandemic.
This phenomenon points to the broader implications of how deeply held beliefs and fears, particularly those amplified by conspiracy theories and misinformation, can mobilize political support for parties that might otherwise remain on the fringes of national politics.
However, the study is not without its limitations. The focus on Japan raises questions about the generalizability of the findings to other contexts, where the political landscape and public attitudes toward vaccines may differ. Additionally, the definition of anti-vaxxers based on their social media behavior assumes that following certain accounts reflects underlying attitudes, which may not always be the case. Future research could benefit from incorporating survey data on voting behavior and exploring the causal relationships between social media engagement, anti-vaccine attitudes, and political orientation.
“The most challenging aspect of conducting this research was applying machine learning and data analysis techniques to vast feeds of Twitter data that were constantly changing,” said Toriumi. “This was done to classify patterns of people’s attitude changes toward the COVID-19 vaccine, distinguishing between persistent anti-vaxxers and pandemic-induced new anti-vaxxers.”
“In the future, we intend to explore the effectiveness of different communication strategies in addressing vaccine hesitancy and misinformation. Additionally, we plan to investigate the role of social media platforms and their algorithms in amplifying or mitigating the spread of conspiracy theories and misinformation. Understanding these dynamics is crucial for developing effective interventions to promote public health and combat misinformation.”
The study, “Anti-vaccine rabbit hole leads to political representation: the case of Twitter in Japan,” was authored by Fujio Toriumi, Takeshi Sakaki, Tetsuro Kobayashi, and Mitsuo Yoshida.
