An Analysis of People’s Emotional Change Toward Vaccines and Its Factors in the Corona Disaster

Abstract

The developers of new vaccines against SARS-CoV-2 and governments have provided information on vaccine effectiveness and status on a daily basis to reassure people about vaccination against COVID-19. However, because the interest in vaccines and vaccination status varies by country and region, people do not always feel reassured. In this paper, we analyzed tweets posted on Twitter to elucidate the emotions people have toward COVID-19 vaccines and factors that cause such emotions to be expressed. We selected six countries for our analysis: Japan, the United States, the Great Britain, Canada, Australia, and India, and applied an emotion classification method using machine learning based on the eight types of emotions defined in Plutchik’s wheel of emotions. We also used a text analysis approach using dependency analysis and burst detection methods. The results of our emotion classification showed that fear was the most common emotion in Japan whereas anger and disgust were most common in the United States, Great Britain, Canada, and Australia; joy was most common in India. We also analyzed tweets during the period when a particular emotion was increased in the changes of the emotions represented as a time series based on the burst-detected dependency relations, and found several characteristics: many users posted vaccine-related news, one user would often post a large number of tweets with the same content, and the same event related to vaccines could arouse different emotions depending on the individual’s situation.