Code for paper Emotion Regulation and Dynamics of Moral Concerns During the Early COVID-19 Pandemic
In this work, we measure dynamics of emotions and moral concerns during the COVID pandemic.
We analyze English-language messages about the pandemic on Twitter between January and May 2020. Dataset can be found here.
We use state-of-the-art language model SpanEmo to measure emotions, including fear, anger, and optimism. Results are compared to that of a simpler method, keyword matching using EmoLex.
We use the FrameAxis method to assess intensity of moral reactions along dimensions such as care, harm, fairness and cheating.
We also split heterogeneous population of users along the ideological line (method) and separately study each user group's emotional and moral dynamics. We find diverging reactions due to both the pandemic and the ideological differences.
Aggregated emotion and moral foundation data can be found here.