Quantification of cultural practices and diversity: An empirical experiment with generative artificial intelligence

Abstract

Culture is often viewed as a value system that shapes cultural practices. Frameworks like Hofstede, GLOBE, and Schwartz identify and quantify various cultural dimensions; however, these rely on surveys that are criticized for limited country coverage, lack of psychometric robustness, small sample sizes, and cultural biases. This article presents an empirical experiment designed to quantify cultural practices and diversity across 216 countries and territories by prompting large language models using a zero-shot learning strategy. This approach enables subnational and segment-specific analyses, equipping researchers with powerful tools for deeper cultural insights.