Why anti-vax opinion may spread so easily
The World Health Organization (WHO) ranked vaccine hesitancy among the top 10 global health threats(opens in new window) in 2019 – before the COVID-19 pandemic. Reluctance to get vaccinated has long been on the rise in Europe(opens in new window), contributing to outbreaks of preventable diseases such as measles. How exactly do such attitudes evolve? What are the drivers of views on vaccination, and how does pro- and anti-vax opinion spread? The DYNAMOD-VACCINE-DATA(opens in new window) project, undertaken with the support of the Marie Skłodowska-Curie Actions programme(opens in new window), has shed new light on these complex dynamics. Crucially, they found that the part of the studied population with neutral attitudes towards vaccines could be more easily swayed to the anti-vaccine side. Work focused on vaccination data up to 2019, including data on trust in vaccines and vaccination coverage in more than 140 countries. “The neutral population was closer to the anti-vaccination side in terms of opinions. We were able to detect higher potential influence on those with neutral views and predict vaccination behaviour in the following year,” says Dino Carpentras, Marie Skłodowska-Curie fellow and former researcher at the University of Limerick(opens in new window).
Like characters in a video game
To understand this connection and its implications, the DYNAMOD-VACCINE-DATA team, supervised by Mike Quayle, used opinion dynamics models in combination with real-world data to provide evidence of social influence. “Opinion dynamics models have a lot in common with video games: they model systems as composed by characters – called agents – following specific rules. They mimic how people interact with each other, making opinion dynamics a powerful tool for studying social phenomena,” Carpentras explains. The team investigated the impact of using real opinion data in such models, which had so far remained largely unexplored.
Mapping influence patterns
They also developed a method for representing this data as a network, detecting links between attitudes and making predictions on this basis. “In practice, the network method allows us to check if people who selected one answer are likely to select another. This makes it possible to draw connections between answers which were selected by the same people, and between people who selected the same answers,” Carpentras remarks. This is how the team was able to discover that those with neutral attitudes towards vaccination were more connected to the anti-vaxxers: the two groups had more attitudes in common with each other than with the pro-vaccine group. Those with a neutral stance were therefore more likely to be influenced by the anti-vaccine group. These findings do not necessarily capture the dynamics at play during the recent pandemic, Carpentras notes: “Opinions are mostly shaped by people discussing. The models are mainly based on slow and horizontal processes; during the COVID-19 crisis, the rhythm of the discussion massively increased, while political discourse and the media (i.e. vertical processes) played a fundamental role.”
More trust with targeted communication
Going forward, however, the project’s outputs could help take action to foster more trust in vaccination: “The network highlights the connections between the neutral and the anti-vaccine sides in different societies. The shorter this distance, the higher the vulnerability to vaccine hesitancy.” Simulations show how targeting different parts of the population with communication campaigns could affect outcomes in each scenario. Results so far indicate that the method could potentially be applied to other social issues, too. If confirmed by further research, it could provide important clues for predicting and influencing attitudes on questions such as climate change.
