The (mis)use of performance quartiles in metacognition and face perception: A comment on Zhou and Jenkins (2020) and Estudillo and Wong (2021)

Kramer, Robin S. S., McIntosh, Robert D. and Nuhfer, Edward B. (2023) The (mis)use of performance quartiles in metacognition and face perception: A comment on Zhou and Jenkins (2020) and Estudillo and Wong (2021). Psychological Reports . ISSN 0033-2941

Full content URL: https://doi.org/10.1177/00332941231181483

Documents
The (mis)use of performance quartiles in metacognition and face perception: A comment on Zhou and Jenkins (2020) and Estudillo and Wong (2021)
Author's accepted manuscript
[img]
[Download]
[img]
Preview
PDF
manuscript.pdf - Whole Document

520kB
Item Type:Article
Item Status:Live Archive

Abstract

A common measurement convention within the field of metacognition is to divide participants into quartiles based on task performance, and then compare self-estimated and actual scores within these sub-groups. This analysis strategy created the famous Dunning-Kruger effect, which asserts that the poorest performers tend to grossly overestimate their abilities. A study by Zhou and Jenkins (2020) has recently replicated this effect within the domain of face matching. However, it can be shown that the analysis strategy induces numerical artefacts prone to misinterpretation, and that randomly generated data lead to the same pattern of results. Estudillo and Wong (2021) used a different quartiles-based approach to argue that only the lowest and highest performers on a task of face recognition showed some insight into their performance. Again, a numerical artefact can explain their result, with the restricted range of the second and third quartiles causing reduced observed correlations between actual and self-estimated abilities. These studies highlight the need for methodological caution when exploring metacognitive questions, and we outline some avenues that may aid future investigation.

Keywords:metacognition, quartiles, Dunning-Kruger effect, face perception, face matching, regression to the mean
Subjects:C Biological Sciences > C800 Psychology
C Biological Sciences > C850 Cognitive Psychology
C Biological Sciences > C830 Experimental Psychology
Divisions:College of Social Science > School of Psychology
ID Code:54906
Deposited On:11 Jul 2023 13:29

Repository Staff Only: item control page