A robust effect size measure for MANOVA with non-normal and non-homogenous data

A common research question in psychology entails examining whether significant group differences (e.g. male and female) can be found in a list of numeric variables that measure the same underlying construct (e.g. intelligence). Researchers often use a multivariate analysis of variance (MANOVA), whic...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Johnson Ching-Hong Li, Marcello Nesca, Rory Michael Waisman, Yongtian Cheng, Virginia Man Chung Tze
Formato: article
Lenguaje:EN
Publicado: SAGE Publishing 2021
Materias:
H
Acceso en línea:https://doaj.org/article/0d3925886a4c42d98ec4bf18b6922983
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:0d3925886a4c42d98ec4bf18b6922983
record_format dspace
spelling oai:doaj.org-article:0d3925886a4c42d98ec4bf18b69229832021-11-30T23:36:02ZA robust effect size measure for MANOVA with non-normal and non-homogenous data2059-799110.1177/20597991211055949https://doaj.org/article/0d3925886a4c42d98ec4bf18b69229832021-11-01T00:00:00Zhttps://doi.org/10.1177/20597991211055949https://doaj.org/toc/2059-7991A common research question in psychology entails examining whether significant group differences (e.g. male and female) can be found in a list of numeric variables that measure the same underlying construct (e.g. intelligence). Researchers often use a multivariate analysis of variance (MANOVA), which is based on conventional null-hypothesis significance testing (NHST). Recently, a number of quantitative researchers have suggested reporting an effect size measure (ES) in this research scenario because of the perceived shortcomings of NHST. Thus, a number of MANOVA ESs have been proposed (e.g. generalized eta squared η Λ 2 , generalized omega squared ω Λ 2 ), but they rely on two key assumptions—multivariate normality and homogeneity of covariance matrices—which are frequently violated in psychological research. To solve this problem we propose a non-parametric (or assumptions-free) ES ( A w ) for MANOVA. The new ES is developed on the basis of the non-parametric A in ANOVA. To test A w we conducted a Monte-Carlo simulation. The results showed that A w was accurate (robust) across different manipulated conditions—including non-normal distributions, unequal covariance matrices between groups, total sample sizes, sample size ratios, true ES values, and numbers of dependent variables—thereby providing empirical evidence supporting the use of A w , particularly when key assumptions are violated. Implications of the proposed A w for psychological research and other disciplines are also discussed.Johnson Ching-Hong LiMarcello NescaRory Michael WaismanYongtian ChengVirginia Man Chung TzeSAGE PublishingarticleSocial SciencesHENMethodological Innovations, Vol 14 (2021)
institution DOAJ
collection DOAJ
language EN
topic Social Sciences
H
spellingShingle Social Sciences
H
Johnson Ching-Hong Li
Marcello Nesca
Rory Michael Waisman
Yongtian Cheng
Virginia Man Chung Tze
A robust effect size measure for MANOVA with non-normal and non-homogenous data
description A common research question in psychology entails examining whether significant group differences (e.g. male and female) can be found in a list of numeric variables that measure the same underlying construct (e.g. intelligence). Researchers often use a multivariate analysis of variance (MANOVA), which is based on conventional null-hypothesis significance testing (NHST). Recently, a number of quantitative researchers have suggested reporting an effect size measure (ES) in this research scenario because of the perceived shortcomings of NHST. Thus, a number of MANOVA ESs have been proposed (e.g. generalized eta squared η Λ 2 , generalized omega squared ω Λ 2 ), but they rely on two key assumptions—multivariate normality and homogeneity of covariance matrices—which are frequently violated in psychological research. To solve this problem we propose a non-parametric (or assumptions-free) ES ( A w ) for MANOVA. The new ES is developed on the basis of the non-parametric A in ANOVA. To test A w we conducted a Monte-Carlo simulation. The results showed that A w was accurate (robust) across different manipulated conditions—including non-normal distributions, unequal covariance matrices between groups, total sample sizes, sample size ratios, true ES values, and numbers of dependent variables—thereby providing empirical evidence supporting the use of A w , particularly when key assumptions are violated. Implications of the proposed A w for psychological research and other disciplines are also discussed.
format article
author Johnson Ching-Hong Li
Marcello Nesca
Rory Michael Waisman
Yongtian Cheng
Virginia Man Chung Tze
author_facet Johnson Ching-Hong Li
Marcello Nesca
Rory Michael Waisman
Yongtian Cheng
Virginia Man Chung Tze
author_sort Johnson Ching-Hong Li
title A robust effect size measure for MANOVA with non-normal and non-homogenous data
title_short A robust effect size measure for MANOVA with non-normal and non-homogenous data
title_full A robust effect size measure for MANOVA with non-normal and non-homogenous data
title_fullStr A robust effect size measure for MANOVA with non-normal and non-homogenous data
title_full_unstemmed A robust effect size measure for MANOVA with non-normal and non-homogenous data
title_sort robust effect size measure for manova with non-normal and non-homogenous data
publisher SAGE Publishing
publishDate 2021
url https://doaj.org/article/0d3925886a4c42d98ec4bf18b6922983
work_keys_str_mv AT johnsonchinghongli arobusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT marcellonesca arobusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT rorymichaelwaisman arobusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT yongtiancheng arobusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT virginiamanchungtze arobusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT johnsonchinghongli robusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT marcellonesca robusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT rorymichaelwaisman robusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT yongtiancheng robusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
AT virginiamanchungtze robusteffectsizemeasureformanovawithnonnormalandnonhomogenousdata
_version_ 1718406184726167552