SENTIMENT ANALYSIS OF COVID-19 VACCINATION POSTS ON FACEBOOK IN INDONESIA WITH CROWDTANGLE

Abstract

COVID-19 was declared a pandemic by the World Health Organization (WHO) due to the virus's uncontrolled spread around the world, including in Indonesia. The Indonesian government has made a number of steps to stop COVID-19 from spreading, one of which is COVID-19 vaccinations. However, not everyone thinks the vaccination is a good idea. Just like in many other countries, Indonesian people responded in different ways to COVID-19 vaccination posts on social media, be it from government officials/agencies or news portals. Their responses can be used to help the government decide on a better vaccination strategy that will help minimize the virus spread and end the pandemic in Indonesia. Using the lexicon method in determining the content sentiment in COVID-19 vaccination posts on Facebook, this research found that unlike news portals that tend to post a more balanced content (36% positive, 20% negative, and 44% neutral out of 23,623 posts with min score = -19, max score = 24, mean score = 0.25, and SD = = 1.43), government accounts posted much more positive content, both in terms of quality (min score = -15, max score = 40, mean score = 4.16, SD = 6.76 ) and quantity (69% classified as positive) than they did the neutral (15%) and the negative content (16%) out of 723 posts. Subsequent analysis with Two-Way ANOVA tests discovered that COVID-19 vaccination posts by the news portal accounts can elicit more varied reactions from the public than government accounts that tended to elicit only positive reactions. It is also confirmed that both the content sentiment of COVID-19 vaccination posts in Indonesia and the account types making the posts as well as their interaction terms do have an impact on how the public responses to them.