Sentiment and Moral Narratives during COVID-19

Oscar Araque, Kyriaki Kalimeri, Lorenzo Gatti

Keywords: covid19, moral foundation theory, sentiment analysis, emotion analysis, natural language processing, twitter

TL;DR: Analysis of both emotion and moral values from Twitter messages during the COVID19 crisis, stuying their relation to misinformation.

Abstract: The COVID-19 pandemic radically changed the lives of millions with unprecedented economic, societal, but also psychological consequences. Containing the pandemic urged for rapid measures, the success of which hugely rely on mass cooperation. Here, we employed a sentiment (DepechMood$++$) and a morality (MoralStrenght) dictionary, to analyse user-generated text on the Twitter platform. Quantifying cooperation and competitive attitudes from the linguistic cues during the pandemic, we depict the moral profiles associated with each position. Moreover, we discuss the moral and emotional profiles of individuals who tweeted about conspiracy theories, and the evolution of moral values throughout the pandemic period for four case-study countries. Our findings show the evolution of moral values during the pandemic for Italy, United States, New Zealand and China. Comparing the moral profiles of cooperative versus competitive attitudes, we find that high loyalty and consideration of the group members is positively correlated to the sentiment of inspiration and negatively correlated to fear.