486. Understanding Public Perception of COVID-19 Social Distancing on Twitter

BACKGROUND: Managing and changing public opinion and behavior are vital for social distancing to successfully slow transmission of COVID-19, preserve hospital resources, and prevent overwhelming the healthcare system’s resources. We sought to leveraging organic, large-scale discussion on Twitter about social distancing to understand public’s beliefs and opinions on this policy. METHODS: Between March 27 and April 10, 2020, we sampled 574,903 English tweets that matched the two most trending social distancing hashtags at the time, #socialdistancing and #stayathome. We used natural language processing techniques to conduct a sentiment analysis that identifies tweet polarity and emotions. We also evaluated the subjectivity of tweets and estimated the frequency of discussion of social distancing rules. We then identified clusters of discussion using topic modeling and compared the sentiment by topic. RESULTS: There was net positive sentiment toward both #socialdistancing and #stayathome with mean sentiment scores of 0.150 (standard deviation [SD], 0.292) and 0.144 (SD, 0.287) respectively. Tweets were also more likely to be objective (median, 0.40; IQR, 0 to 0.6) with approximately 30% of all tweets labeled as completely objective. Approximately half (50.4%) of all tweets primarily expressed joy and one-fifth expressed fear and surprise each (Figure 1). These trends correlated well with topic clusters identified by frequency including leisure activities and community support (i.e., joy), concerns about food insecurity and effects of the quarantine (i.e., fear), and unpredictability of COVID and its unforeseen implications (i.e., surprise) (Table 1). Table 1. Topic clusters identified by topic modeling. Words contributing to the model are shown in decreasing order of weighting. The topics are labeled manually based on these words. The number of tweets primarily with that topic, mean sentiment, mean subjectivity, and sample tweets are also included. [Image: see text] Figure 1. Emotion analysis for all tweets and stratified by tweets with the hashtag #socialdistancing and #stayathome. Comparison between the two hashtags is done using Chi-squared testing. Bonferroni correction was used to define statistical significance at a threshold of p = 0.008 (0.05/n, where n = 6 since 6 comparisons were completed). [Image: see text] CONCLUSION: The positive sentiment, preponderance of objective tweets, and topics supporting coping mechanisms led us to believe that Twitter users generally supported social distancing measures in the early stages of their implementation. DISCLOSURES: All Authors: No reported disclosures