Zobraziť v EDS

The receiver operating characteristic area under the curve (or mean ridit) as an effect size.

Uložené v:
Podrobná bibliografia
Názov: The receiver operating characteristic area under the curve (or mean ridit) as an effect size.
Autori: Smithson M; Research School of Psychology, Australian National University.
Zdroj: Psychological methods [Psychol Methods] 2025 Jun; Vol. 30 (3), pp. 678-686. Date of Electronic Publication: 2023 Jul 13.
Spôsob vydávania: Journal Article
Jazyk: English
Informácie o časopise: Publisher: American Psychological Association Country of Publication: United States NLM ID: 9606928 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1939-1463 (Electronic) Linking ISSN: 1082989X NLM ISO Abbreviation: Psychol Methods Subsets: MEDLINE
Imprint Name(s): Original Publication: Washington, DC : American Psychological Association, c1996-
Výrazy zo slovníka MeSH: ROC Curve* , Area Under Curve* , Psychology*/methods, Humans ; Data Interpretation, Statistical
Abstrakt: Several authors have recommended adopting the receiver operator characteristic (ROC) area under the curve (AUC) or mean ridit as an effect size, arguing that it measures an important and interpretable type of effect that conventional effect-size measures do not. It is base-rate insensitive, robust to outliers, and invariant under order-preserving transformations. However, applications have been limited to group comparisons, and usually just two groups, in line with the popular interpretation of the AUC as measuring the probability that a randomly chosen case from one group will score higher on the dependent variable than a randomly chosen case from another group. This tutorial article shows that the AUC can be used as an effect size for both categorical and continuous predictors in a wide variety of general linear models, whose dependent variables may be ordinal, interval, or ratio level. Thus, the AUC is a general effect-size measure. Demonstrations in this article include linear regression, ordinal logistic regression, gamma regression, and beta regression. The online supplemental materials to this tutorial provide a survey of currently available software resources in R for the AUC and ridits, along with the code and access to the data used in the examples. (PsycInfo Database Record (c) 2025 APA, all rights reserved).
Entry Date(s): Date Created: 20230713 Date Completed: 20250522 Latest Revision: 20250529
Update Code: 20250529
DOI: 10.1037/met0000601
PMID: 37439716
Databáza: MEDLINE
Popis
Abstrakt:Several authors have recommended adopting the receiver operator characteristic (ROC) area under the curve (AUC) or mean ridit as an effect size, arguing that it measures an important and interpretable type of effect that conventional effect-size measures do not. It is base-rate insensitive, robust to outliers, and invariant under order-preserving transformations. However, applications have been limited to group comparisons, and usually just two groups, in line with the popular interpretation of the AUC as measuring the probability that a randomly chosen case from one group will score higher on the dependent variable than a randomly chosen case from another group. This tutorial article shows that the AUC can be used as an effect size for both categorical and continuous predictors in a wide variety of general linear models, whose dependent variables may be ordinal, interval, or ratio level. Thus, the AUC is a general effect-size measure. Demonstrations in this article include linear regression, ordinal logistic regression, gamma regression, and beta regression. The online supplemental materials to this tutorial provide a survey of currently available software resources in R for the AUC and ridits, along with the code and access to the data used in the examples. (PsycInfo Database Record (c) 2025 APA, all rights reserved).
ISSN:1939-1463
DOI:10.1037/met0000601