Understanding coronavirus disease 2019 (COVID-19) vaccine hesitancy: Evidence from the community-driven knowledge site Quora.

Objective: The present study aims to examine the threshold of coronavirus disease 2019 (COVID-19) vaccine hesitancy over time and public discourse around COVID-19 vaccination hesitancy. Methods: We collected 3,952 questions and 66,820 answers regarding COVID-19 vaccination posted on the social question-and-answer website Quora between June 2020 and June 2021 and employed Word2Vec and Sentiment Analysis to analyze the data. To examine changes in the perceptions and hesitancy about the COVID-19 vaccine, we segmented the data into 25 bi-weekly sections. Results: As positive sentiment about vaccination increased, the number of new vaccinations in the United States also increased until it reached a ceiling point. The vaccine hesitancy phase was identified by the decrease in positive sentiment from its highest peak. Words that occurred only when the positive answer rate peaked (e.g., safe, plan, best, able, help) helped explain factors associated with positive perceptions toward vaccines, and the words that occurred only when the negative answer rate peaked (e.g., early, variant, scientists, mutations, effectiveness) suggested factors associated with vaccine hesitancy. We also identified a period of vaccine resistance, where people who decided not to be vaccinated were unlikely to be vaccinated without further enforcement or incentive. Conclusions: Findings suggest that vaccine hesitancy occurred because concerns about vaccine safety were high due to a perceived lack of scientific evidence and public trust in healthcare authorities has been seriously undermined. Considering that vaccine-related conspiracy theories and fake news prevailed in the absence of reliable information sources, restoring public trust in healthcare leaders will be critical for future vaccination efforts.

Anti-Asian Discourse in Quora: Comparison of Before and During the COVID-19 Pandemic with Machine- and Deep-Learning Approaches

The current study attempts to compare anti-Asian discourse before and during the COVID-19 pandemic by analyzing big data on Quora, one of the most frequently used community-driven knowledge sites. We created two datasets regarding “Asians” and “anti-Asians” from Quora questions and answers between 2010 and 2021. A total of 1,477 questions and 5,346 answers were analyzed, and the datasets were divided into two time periods: before and during the COVID-19 pandemic. We conducted machine-learning-based topic modeling and deep-learning-based word embedding (Word2Vec). Before the pandemic, the topics of physical difference and racism were prevalent, whereas, after the pandemic, the topics of hate crime, the need to stop Asian hate crimes, and the need for the Asian solidarity movement emerged. Above all, the semantic similarity between Asian and Black people became closer, while the similarity between Asian people and other racial/ethnic groups was diminished. The emergence of negative and radical language, which increased saliently after the outbreak of the pandemic, and the considerably wider semantic distance between Asian and White people indicates that the relationship between the two races has been weakened. The findings suggest a long-term campaign or education system to reduce racial tensions during the pandemic.