Survey will open in a new window or redirect momentarily, please keep this tab open! If the survey does not open, please enable popups and refresh the page.

Status:

Dismiss

Overconfidence in people and Large Language Models

Study ID sspp-2026-0049-v1

General Details

Project UFS
Study ID sspp-2026-0049-v1
Study Title Overconfidence in people and Large Language Models
Authors Don Moore
Completion Time 15 Minutes
Close Date (UTC) Nov. 30, 2026
Discipline Psychology
Field Social Psychology
Country United States
Abstract
Both people and LLMs completed the same task: guessing others' survey responses, in the style of Family Feud. They then estimated their own and others' performance on this task. These estimates yield measures of overestimation, overplacement, and overprecision. Who was more overconfident by these measures--humans or LLMs?




Citation
Don Moore. 2026. "Overconfidence in people and Large Language Models." Social Science Prediction Platform. October 29. https://socialscienceprediction.org/s/azac82