Expressed Sentiment on Social Media During the COVID-19 Pandemic: Evidence from the Lockdown in Shanghai

The outbreak of the COVID-19 pandemic caused numerous lockdowns in cities around the world. While lockdowns can generally be effective at reducing the spread of COVID-19, they also have substantial psychosocial impacts, such as causing distress and anxiety among residents. However, our knowledge of the effects of lockdown on well-being is limited; for example, we know little about the sentiment alterations of residents during a long-lasting and stringent lockdown. To investigate this issue, we assemble a dataset of about 7 million geotagged microblog tweets from 100,000 randomly selected users on the most popular microblog site in China around the time of the COVID-19 lockdown in Shanghai from February to May 2022. We find that the expressed sentiment of Shanghai residents drops significantly one week into the lockdown. By contrast, the expressed sentiment of residents outside Shanghai rises sharply following the lockdown as it prevents the spread of the COVID-19 infections from Shanghai to other provinces. The expressed sentiment of pessimistic residents, women, and residents with higher education experience a larger drop during the lockdown.