Welcome to the GitHub repository for "Not Yet: Large Language Models Cannot Replace Human Respondents for Psychometric Research". This repository is structured to facilitate an understanding and exploration of the research presented in our article.
-
LLM_Personality_Responses_Generation: Code for generating all LLM data we used in the study.
-
BFI_2_Analyses: Analyses of the Big Five Inventory-2 data.
-
HEXACO_Analyses: Analysis of the HEXACO-100 Personality Inventory data.
-
Additional Analyses_Distinguishing Human vs. LLM Responses: Online supplementary materials for "Additional Analyses: Distinguishing Human vs. LLM Responses", include code and write-up.
-
Additional Analyses_Social Desirability Rating and LLM Responses: Online supplementary materials for "Additional Analyses: Social Desirability Rating and LLM Responses", include code and write-up.
Please see the corresponding analysis files (BFI_2_Analyses and HEXACO_Analyses) for all data and LLM_Personality_Responses_Generation for how to generate LLM responses.