On the 21st of March 2025 (09:30-14:00), the ODISSEI SoDa team is organising a workshop focused on the usage of Large Language Models in the Social Sciences.
With the advent of large language models (LLMs), collecting measurements related to social science constructs (e.g., personality traits, political attitudes, human values) has become easier, faster and more affordable. These measurements are subsequently used for modelling societal and group processes that social scientists typically engage in, where inferences from samples to populations are also made. Valid modelling and inferences (e.g., unbiased estimates, p-values), however, require high-quality measurements or at the very least, methods to deal with the presence of measurement error. Just like traditional questionnaire-based measurements, LLM-based measurements have been shown to suffer from validity and reliability issues. This LLM workshop, offered by the SoDa team, will focus on dealing with measurement errors in LLM-based annotations and correct statistical inferences with variables derived from LLM annotations. It is aimed at (computational) social scientists who are interested in using LLM-based variables for inferential statistical modelling. This will be a hands-on workshop; please take your laptop with you.
Lunch and refreshments will be provided.