A Comparative Evaluation of Structural Topic Models and BERTopic for Short, Open-Ended Survey Responses
A new paper compares two topic modeling approaches, Structural Topic Models (STM) and BERTopic, for analyzing short, open-ended survey responses. The study found that BERTopic generally produced more coherent and interpretable topics, especially when using a novel contextual augmentation strategy. While STM offers stronger inferential capabilities for covariate analysis, BERTopic provides more descriptive comparisons, suggesting both methods have complementary strengths for social science research. AI
IMPACT Provides guidance on selecting and combining topic modeling approaches for analyzing short, open-ended survey data.