Managing Dissent: Institutional Control and Political Discourse Polarisation in Uganda
研究乌干达混合政权如何通过不同制度控制程度的平台（议会、电视、社交媒体）管理政治话语，发现控制越强极化越低，政权并非压制异议而是疏导它。
This study examines how hybrid regimes maintain power by managing political discourse, using Uganda’s 2021 election as a case study. We analyse communication patterns across three distinct settings that vary in institutional control: Parliament (high control), television shows (medium control), and social media platform X (low control). Our dataset comprises 4945 social media posts, 104 parliamentary transcripts, and 157 television episodes collected from May 2020 to September 2021. We focus on four key political topics: Covid-19, the economy, political violence, and elections. Using computational textual analysis, we measure discursive polarisation between government and opposition actors through two metrics: semantic divergence (linguistic distance between actors) and emotional tone. Our findings show a clear institutional gradient in political polarisation. Polarisation is highest on X, where institutional control is minimal, and lowest in Parliament, where control is strongest. This pattern persists across all four topics and shows temporal variations linked to the election cycle. These results suggest that hybrid regimes do not uniformly suppress dissent. Instead, they calibrate their control strategies across different communicative platforms, channelling rather than eliminating opposition discourse. This finding refines theories of informational autocracy by showing how regimes strategically manage political communication across institutionally varied settings.