This research is intended to answer the question of the discourses developed in association with Covid-19 management. Discourse in this regard is understood within the context of public policy, i.e. a set of concepts or ideas conveyed by actors to influence public policies. This study employed a discourse coalition theory. In addition, this study applied a discourse network analysis (DNA) method, combining discourse analysis (qualitative contents) and social network method. The DNA method maps any developing discourses on an issue (what) and also actors stating such discourses (who). The research was carried out for five months, starting from 17th November 2019 to 30th April 2020 and it included 1,123 statements of actors that were published in the media. The results of this study show that a coalition of discourse also occurs in health issues. The study reinforces the findings of Leifeld and Haunss (2011) on how actors attempt to present discourse to dominate the public conversation. This research shows the efforts of government actors, i.e. central government, regional governments, and the Ministry of Health, to conduct a coalition between two opposing discourses. The government efforts are not entirely successful. The government is also less successful in using communication channels and social media to form a single discourse.

This study is intended to answer the question of how political polarization is related to social media users’ posts about Covid-19. The researchers chose health cases related to Covid-19 instead of political issues (e.g. elections) to prove that this political polarization has spread to many areas. The research also wants to see the relationship between this political polarization and selective exposure. Theories applied in this study are polarization, filter bubble, and selective exposure. The study applied two methods: social media network analysis and content analysis. The network analysis included 82,156 posts, while the content analysis was carried out on 4,050 social media accounts. The research outcome proves the occurrence of political polarization. Social media users were divided into two major groups, namely pro-Jokowi and anti-Jokowi. Each group interacted with fellow users who had the same political choices and shared the same message content. Users with certain political choices tend to receive the same information as their political choices, and ignore information from other political parties. Another interesting finding from this study is how this polarization was sharpened by the use of hashtags. Each party (supporters and oppositions of Jokowi) uses hashtags to create solidarity and mobilization from each supporter. Research also proves the validity of the selective exposure and filter bubble hypothesis in the Indonesian context.