Vaccination Uptake, Happiness and Emotions: Using a Supervised Machine Learning Approach

This chapter presents a retrospective evaluation of the COVID-19 pandemic, focusing on the factors influencing vaccination uptake to achieve herd immunity thresholds of 70% and 90% in ten countries (refer to Chaps. 4 , 5 , and 7 through 9 ). We identify the most important variables contributing to high vaccination rates by utilising supervised machine learning techniques, particularly the Extreme Gradient Boosting (XGBoost) algorithm. Our analysis explores the relationship of factors within a global health crisis context by drawing from comprehensive datasets encompassing COVID-19-related information, population demographics, and sentiment analysis from social media. Special consideration is given to exploring whether subjective well-being measures played a role in the decision to get vaccinated since we know from Chap. 9 that negative emotions, such as fear of vaccine side effects, can sway people’s attitudes towards vaccination and that happier people make better health-related decisions. Our findings offer several significant contributions to the literature. Firstly, we conduct a pioneering post-COVID-19 cross-country analysis, identifying factors important for reaching different herd immunity levels. Secondly, we integrate subjective well-being measures into our estimations, providing insights into the influence of emotional states on vaccine uptake. Thirdly, we differentiate between factors impacting the achievement of various vaccination thresholds, shedding light on nuanced dynamics. The XGBoost model is the most effective, delivering precise predictions and highlighting key factors such as vaccination policies, international travel controls, rural population percentages, and average temperature. Interestingly, happiness emerges as an important factor in attaining the 90% vaccination threshold, underscoring the importance of addressing emotional perceptions in vaccination strategies. However, our study acknowledges limitations, including the focus on primarily developed countries and the inability to incorporate variables reflecting international support due to high missingness. Nonetheless, our insights offer actionable recommendations for policymakers to enhance vaccination rates and mitigate pandemics' health, economic, and political impacts.