Computational Archival Analysis in Studying International Conflict

• In Event: Text-as-Data, Archives, and Diplomacy

Abstract

Many puzzles in international security involve decision-making by elite policymakers confronting extraordinary circumstances behind closed doors. This has limited analyses to formal models, elite-based surveys, coarse statistics, and in particular, qualitative case studies using archival documents. I propose that quantitative scholars can also use declassified archival documents to address many of these questions. Computational archival analysis combines features of quantitative, qualitative, and historical approaches, allowing scholars to detail a single conflict period in a systematic and comprehensive manner. After outlining this process and some best practices, I demonstrate these techniques by analyzing daily-level military data and negotiation transcripts from the Korean War of 1950-1953. I show that even though higher costs from fighting are ultimately linked to conflict termination, they generally lead to decreased interest in a negotiated agreement. These results challenge existing scholarship on war termination and highlight the importance of expanding the tools and materials we use to study international affairs.

Author

• Eric Min, UCLA