Author Type

Graduate Student

Date of Award

Spring 4-27-2026

Document Type

Thesis

Publication Status

Version of Record

Submission Date

May 2026

Department

Psychology

College Granting Degree

Charles E. Schmidt College of Science

Department Granting Degree

Psychology

Degree Name

Master of Science (MS)

Thesis/Dissertation Advisor [Chair]

Robin Vallacher

Abstract

Silicon participants are large language model generated research participants who are embedded with demographic information, as well as personality measures. The use of silicon participants in the behavioral sciences field is underexplored. This current study generated silicon participants based on human data (N=514), where each silicon participant matches their human counterpart. Both samples were asked to respond to seven decision-making dilemmas that are written from scratch. The dilemmas differ in their severity and represent the ambiguity of social life. Responses of the silicon participants were generated using ChatGPT 4.0. This is the first study that embedded personality traits and used original dilemmas that are not available on the internet. The results suggest that while silicon participants can replicate some aspects of decision-making observed in humans, further alignment is needed to ensure silicon participants can replicate the nuances of human behavior.

Download

Included in

Psychology Commons

Share

COinS