Large Language Models for Social Science Research

About

The course is organised around four main components. It introduces the basic concepts behind LLMs, including how they are trained, how they generate language, and how they differ from earlier computational text‑analysis approaches. It then explores recent applications in the social sciences, such as automated content coding, large‑scale text analysis, survey augmentation, and experimental design.

You will work through guided demonstrations using simple coding environments with exercises covering prompt‑based classification, document summarisation, semantic search, and basic data extraction.

The course also addresses validation, bias, and responsible use, and discusses how LLMs can be integrated into research workflows.

Course leader

Giovanni Pagano is a Research Fellow in the Department of Social and Political Sciences at the University of Milan. His research centres on political communication, party competition, and computational methods in political science, with a focus on the analysis of texts, visuals, and social media data. Giovanni's work has appeared in journals including European Union Politics, Regulation & Governance, and Journal of Information Technology & Politics. He teaches courses and workshops on quantitative methods and computational approaches for social scientists.

Target group

Social science researchers

Course aim

To provide social science researchers with a practical and conceptual introduction to Large Language Models (LLMs), enabling them to understand how these tools work, critically evaluate their use, and apply them responsibly within their own research workflows.

Fee info