The Measurement of COVID-19 Vaccine Hesitancy

The content presented in this chapter to Chap. 9 comes from our published research in PLOS ONE. This chapter serves as the initial examination into measuring COVID-19 vaccine hesitancy, laying the groundwork for a comprehensive cross-country panel analysis utilising Big Data in Chaps. 8 and 9 . By providing an in-depth literature review on vaccine hesitancy and detailing the methodology employed in data acquisition and analysis, this chapter explains the complexities of gauging attitudes towards COVID-19 vaccines using real-time Twitter data. The chapter focuses on studies utilising Big Data analytics to understand vaccine hesitancy. It explains the process of constructing time-series data by extracting tweets and employing Natural Language Processing techniques to determine sentiment and emotions surrounding COVID-19 vaccines. The chapter introduces key lexicons utilised in sentiment analysis and outlines the methodology for topic modelling and word cloud visualisation. Central to the chapter is the detailed explanation of data acquisition and processing methodologies, wherein over a million tweets are extracted, translated, and analysed to derive sentiment and emotion time series. Robustness tests, including frequency and volume analyses, validate the reliability of the derived time-series data, ensuring the accuracy of subsequent analyses. Moreover, the chapter discusses the importance of topic modelling and word clouds in identifying trends and narratives within the tweet corpus, enhancing our understanding of attitudes towards COVID-19 vaccines. By employing advanced analytical techniques, such as the Latent Dirichlet Allocation model, the chapter aims to organise unstructured text data into meaningful themes, providing valuable insights for policymakers and researchers alike.