How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test

Background: Advances in artificial intelligence have enabled the simulation of human-like behaviors, raising the possibility of using large language models (LLMs) to generate synthetic population samples for research purposes, which may be particularly useful in health and social sciences. Methods:...

Full description

Saved in:
Bibliographic Details
Main Authors: Gregorio Ferreira, Jacopo Amidei, Rubén Nieto, Andreas Kaltenbrunner
Format: Article
Language:English
Published: American Association for the Advancement of Science (AAAS) 2025-01-01
Series:Health Data Science
Online Access:https://spj.science.org/doi/10.34133/hds.0284
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849427473169645568
author Gregorio Ferreira
Jacopo Amidei
Rubén Nieto
Andreas Kaltenbrunner
author_facet Gregorio Ferreira
Jacopo Amidei
Rubén Nieto
Andreas Kaltenbrunner
author_sort Gregorio Ferreira
collection DOAJ
description Background: Advances in artificial intelligence have enabled the simulation of human-like behaviors, raising the possibility of using large language models (LLMs) to generate synthetic population samples for research purposes, which may be particularly useful in health and social sciences. Methods: This paper explores the potential of LLMs to simulate population samples mirroring real ones, as well as the feasibility of using personality questionnaires to assess the personality of LLMs. To advance in that direction, 2 experiments were conducted with GPT-4o using the Eysenck Personality Questionnaire Revised-Abbreviated (EPQR-A) in 6 languages: Spanish, English, Slovak, Hebrew, Portuguese, and Turkish. Results: We find that GPT-4o exhibits distinct personality traits, which vary based on parameter settings and the language of the questionnaire. While the model shows promising trends in reflecting certain personality traits and differences across gender and academic fields, discrepancies between the synthetic populations’ responses and those from real populations remain. Conclusions: These inconsistencies suggest that creating fully reliable synthetic population samples for questionnaire testing is still an open challenge. Further research is required to better align synthetic and real population behaviors.
format Article
id doaj-art-17b1471bdf344b6cbc7e4b43a1295140
institution Kabale University
issn 2765-8783
language English
publishDate 2025-01-01
publisher American Association for the Advancement of Science (AAAS)
record_format Article
series Health Data Science
spelling doaj-art-17b1471bdf344b6cbc7e4b43a12951402025-08-20T03:29:02ZengAmerican Association for the Advancement of Science (AAAS)Health Data Science2765-87832025-01-01510.34133/hds.0284How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality TestGregorio Ferreira0Jacopo Amidei1Rubén Nieto2Andreas Kaltenbrunner3IN3, Universitat Oberta de Catalunya, Barcelona, Spain.IN3, Universitat Oberta de Catalunya, Barcelona, Spain.eHealth Research Lab, Universitat Oberta de Catalunya, Barcelona, Spain.IN3, Universitat Oberta de Catalunya, Barcelona, Spain.Background: Advances in artificial intelligence have enabled the simulation of human-like behaviors, raising the possibility of using large language models (LLMs) to generate synthetic population samples for research purposes, which may be particularly useful in health and social sciences. Methods: This paper explores the potential of LLMs to simulate population samples mirroring real ones, as well as the feasibility of using personality questionnaires to assess the personality of LLMs. To advance in that direction, 2 experiments were conducted with GPT-4o using the Eysenck Personality Questionnaire Revised-Abbreviated (EPQR-A) in 6 languages: Spanish, English, Slovak, Hebrew, Portuguese, and Turkish. Results: We find that GPT-4o exhibits distinct personality traits, which vary based on parameter settings and the language of the questionnaire. While the model shows promising trends in reflecting certain personality traits and differences across gender and academic fields, discrepancies between the synthetic populations’ responses and those from real populations remain. Conclusions: These inconsistencies suggest that creating fully reliable synthetic population samples for questionnaire testing is still an open challenge. Further research is required to better align synthetic and real population behaviors.https://spj.science.org/doi/10.34133/hds.0284
spellingShingle Gregorio Ferreira
Jacopo Amidei
Rubén Nieto
Andreas Kaltenbrunner
How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test
Health Data Science
title How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test
title_full How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test
title_fullStr How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test
title_full_unstemmed How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test
title_short How Well Do Simulated Population Samples with GPT-4 Align with Real Ones? The Case of the Eysenck Personality Questionnaire Revised-Abbreviated Personality Test
title_sort how well do simulated population samples with gpt 4 align with real ones the case of the eysenck personality questionnaire revised abbreviated personality test
url https://spj.science.org/doi/10.34133/hds.0284
work_keys_str_mv AT gregorioferreira howwelldosimulatedpopulationsampleswithgpt4alignwithrealonesthecaseoftheeysenckpersonalityquestionnairerevisedabbreviatedpersonalitytest
AT jacopoamidei howwelldosimulatedpopulationsampleswithgpt4alignwithrealonesthecaseoftheeysenckpersonalityquestionnairerevisedabbreviatedpersonalitytest
AT rubennieto howwelldosimulatedpopulationsampleswithgpt4alignwithrealonesthecaseoftheeysenckpersonalityquestionnairerevisedabbreviatedpersonalitytest
AT andreaskaltenbrunner howwelldosimulatedpopulationsampleswithgpt4alignwithrealonesthecaseoftheeysenckpersonalityquestionnairerevisedabbreviatedpersonalitytest