Research

This thesis examines the effect of threat perception on protest repression by focusing on protest and repression dynamics in Turkey during democratic backsliding. Questioning when and against whom protest repression is more intense, it seeks to understand the determinants of state response to dissent. However, traditional media sourcing and news outlets’ coverage are insufficient in non-democratic settings where the media is co-opted, censored, and heavily repressed, which requires addressing additional causes of selection bias. To address this problem, I construct and leverage a unique dataset based on human rights reports. The TRACE dataset measures protest and repression events in Turkey since 2013 with fine-grained event categorization and issue type identification. The comparison with other protest event datasets illustrates TRACE’s strength in measuring protest repression across different issues in Turkey. To analyze changes in protest repression behavior across various issue types, I employ multinomial logistic regression. Results show that the intensity of protest repression increases with the perceived threat. Further, mobilization becomes more threatening to the regime when it is challenged by labor and environmental issues. Ultimately, this thesis aims to provide more insight into targeted protest repression and threat perceptions in non-democratic regimes.