Browsing by Subject "Computational text analysis"

Now showing 1 - 1 of 1

Open Access
Economic hardships, polarizing rhetoric, and diversionary use of force
(2021-05) Yılmaz, Burak Kazim
Experiencing economic and social problems make state leaders lose popularity among their constituencies. Eventually, they are likely to lose their positions, if problems persist. Motivated to stay in office, leaders resort to different policies as long as political systems that they exist within allow. Use of force to divert public attention is one of these policies. Scholars who have studied this phenomenon suggested several mechanisms to trace state leaders' efforts for creating diversionary conflict. In this thesis, I offer a novel approach which adds socities' political polarization into the relationship between leaders' need for diversion and conflict initiation. To test this approach, I chose post-2014 presidential election Turkey as my case study. The case is rich in content because of the deteriorating economic indicators, presence of a polarized society, and having different violent conflicts. I measured the polarization variable through president's speeches by using computational text analysis methods. Then, I conducted regression analyses to detect the relationship between polarizing rhetoric and diversionary use of force. My findings reveal that the ruling party elite has followed different practices in using polarizing elements in their speeches for conflicts inside and outside Turkey. While rhetoric becomes more polarized in the times that armed forces conduct more operations to PKK, it becomes less when the number of conflicts abroad increases.