EU citizens and intention to get vaccinated: a Bayesian multilevel analysis

Abstract Vaccine hesitancy is a serious threat during a pandemic because it slows down the increase of vaccination coverage in the population. The study of individuals’ attitudes towards vaccines before and during the COVID-19 pandemic is not a novelty, but here we investigate the determinants of the propensity towards vaccination among the European citizens living in 80 geographical areas belonging to the EU-27 countries. Given the multilevel structure of data with citizens nested within NUTS-1 geographical areas, we select a Bayesian hierarchical model to estimate the effect of several variables of trust measured at individual and at NUTS-1 level (i.e., trust in news and social media, police, Government, health care system, science, pharmaceutical companies), plus other controls, on the propensity towards vaccination. The study relies on data of the Eurofound survey (Round 3) “Living, working, and COVID-19”, collected at the beginning of the COVID-19 vaccine campaign (Spring 2021) and involving more than 45,000 individuals across the Member States of the European Union. Our analysis allows us to predict the probability to get vaccinated for different profiles of individuals that are distinguished for individual and NUTS-1 level characteristics, thus providing policy makers and opinion leaders with a useful tool to identify the key elements on which to leverage to convince people about the goodness of vaccines.