Using Large Language Models in Political Science

Abstract

Recent advancements in Natural Language Processing (NLP) and the emergence of powerful Large Language Models (LLMs) such as OpenAI’s ChatGPT, DepSeek R1, Meta LLAMA, Mistral have opened new avenues for research in political science. LLMs made methods that were technically possible but time consuming, such as automated content analysis of political speeches and party manifestos, accessible. Unimaginable methods – such as synthetic surveys (silicone samples) and synthetic experiments - are now reality. Some integrations can even generate an entire research paper from scratch, provided only with a research question and hypothesis. This course focuses on applying modern LLM-based techniques in political science research and offers a structured overview. In 4 weeks, we will cover the foundations of NLP, explore applications, employ LLM-based analysis, and learn how to assess the reliability and validity of these methods critically. By the end of the course, students will be equipped with both the theoretical understanding and practical tools to incorporate the LLM-driven NLP into their own research on political phenomena.