Knowing Your Uncertainty -- On the application of LLM in social sciences
This work addresses the problem of ensuring rigorous and reliable use of LLMs in computational social science research, which is incremental as it builds on existing uncertainty quantification methods from both computer science and social sciences.
The paper tackles the challenge of applying large language models (LLMs) to social science research by emphasizing the need for uncertainty assessment, introducing a unified framework that categorizes tasks and validation types to map existing uncertainty quantification methods and offer practical recommendations.
Large language models (LLMs) are rapidly being integrated into computational social science research, yet their blackboxed training and designed stochastic elements in inference pose unique challenges for scientific inquiry. This article argues that applying LLMs to social scientific tasks requires explicit assessment of uncertainty-an expectation long established in both quantitative methodology in the social sciences and machine learning. We introduce a unified framework for evaluating LLM uncertainty along two dimensions: the task type (T), which distinguishes between classification, short-form, and long-form generation, and the validation type (V), which captures the availability of reference data or evaluative criteria. Drawing from both computer science and social science literature, we map existing uncertainty quantification (UQ) methods to this T-V typology and offer practical recommendations for researchers. Our framework provides both a methodological safeguard and a practical guide for integrating LLMs into rigorous social science research.