Automated event data extraction techniques have revolutionized the study of conflict dynamics through the ability of these techniques to generate large volumes of timely data measuring dynamic interactions among actors around the world. In this paper, we describe our approach for adapting these techniques to extract data on sentiments and emotions, which are theorized to crucially contribute to escalating and de-escalating conflict. Political scientists view political conflict as resulting from a series of strategic interactions between groups and individuals. Psychologists highlight additional factors in political conflict, such as endorsements and condemnations, the public’s attitude toward its leaders, the impact of public attitudes on policy, and decisions to engage in armed conflict. This project combines these two approaches to examine hypotheses regarding the effects that different emotional impulses have on government and dissident decisions to escalate or de-escalate their use of hostility and violence. Across the two cases examined—the democratic Philippines and authoritative Egypt between 2001 and 2012—we found consistent evidence that intense societal fear of dissidents and societal disgust toward the government were associated with increases in dissident hostility. Conversely, societal anger toward dissidents was associated with a reduction in dissident hostility. However, we also found noticeable differences between the two regimes. We close the article with a summary of these similarities and differences, along with an assessment of their implications for future conflict studies.
Natural language processing automated events and sentiment extraction conflict dynamics emotions and behavior conflict early warning
Primary Language | English |
---|---|
Journal Section | Articles |
Authors | |
Publication Date | August 22, 2013 |
Published in Issue | Year 2013 |
Widening the World of IR