Communicating Concerns, Emotional Expressions, and Disparities on Ethnic Communities on Social Media During the COVID-19 Pandemic: A Structural Topic Modeling Approach

Abstract

Ethnic and racial disparities in the coronavirus (COVID-19) pandemic raise significant concerns. This study analyzes social media discourses toward four ethnic communities in the United States during the pandemic and reveals disparities in pandemic experiences among them. A total of 488,029 tweets mentioning one of the four ethnic communities, that is, Asians, Blacks, Hispanics, and Native Americans, were investigated by a structural topic modeling approach with emotional expressions and time as covariates in the topic model. The results demonstrate that discourses about Asian, Hispanics, and Native American communities were often induced by pandemic-related events, concerning topics beyond one’s community, and reflecting an experience of implicit racism and an adoption of technical supports from health systems. Meanwhile, discourses about Blacks were racially related, discussing topics within the community, and reflecting an experience of explicit racism and an adoption of psychological supports from ingroup. We discuss the implications of our findings on ethnic health disparities.