Shifting Trends of COVID-19 Tweet Sentiment with Respect to Voting Preferences in the 2020 Election Year of the United States
This research addresses how social media sentiment reflects political polarization during a crisis, but it is incremental as it applies existing sentiment analysis methods to new election-year data.
The study analyzed COVID-19 tweet sentiment in relation to U.S. voting preferences during the 2020 election, finding weak correlations with state-level popular votes and a shift from more positive sentiments in blue states early in lockdowns to more positive sentiments in red states by summer and election day.
COVID-19 related policies were extensively politicized during the 2020 election year of the United States, resulting in polarizing viewpoints. Twitter users were particularly engaged during the 2020 election year. Here we investigated whether COVID-19 related tweets were associated with the overall election results at the state level during the period leading up to the election day. We observed weak correlations between the average sentiment of COVID-19 related tweets and popular votes in two-week intervals, and the trends gradually become opposite. We then compared the average sentiments of COVID-19 related tweets between states called in favor of Republican (red states) or Democratic parties (blue states). We found that at the beginning of lockdowns sentiments in the blue states were much more positive than those in the red states. However, sentiments in the red states gradually become more positive during the summer of 2020 and persisted until the election day.