Quantitative Text Analysis in International Relations: A Review and Perspective

This article reviews the development and current landscape of quantitative text analysis in international relations (IR), with particular attention to its applications to United Nations documents. Over the past two decades, the large-scale digitization of diplomatic records and advances in natural language processing (NLP) have transformed how textual data can be analyzed in social science research. The article surveys foundational approaches, including dictionary-based methods, topic modeling, and machine-learning-based classification, and shows how they have been used to study international norms, policy preferences, and diplomatic discourse. It then discusses recent methodological advances associated with deep learning, especially Transformer-based models and large language models (LLMs), and illustrates their analytical potential through applications to United Nations Security Council debates. These examples demonstrate how contemporary NLP enables fine-grained analyses of normative change and agenda dynamics. The article concludes by reflecting on key methodological challenges and future directions for quantitative text analysis in IR.