EUVIMEDEUVIMED entdecken
MetaanalyseOpen Access · CC BY-NC 4.0

A HOMOGENEITY TEST FOR META-ANALYSIS WITH STANDARDIZED MEAN DIFFERENCES

Juan Botella, Manuel Suero, Juan Ignacio Durán

Acción Psicológica · 11.06.2026

DOI: 10.5944/ap.46029 ↗

Kurz gefasst

Metaanalyse in der Zeitschrift Acción Psicológica, online erschienen am 11.06.2026. Verfasst von Juan Botella und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Testing the homogeneity hypothesis (i.e., that data from a set of related studies share a common parametric effect size) is a traditional part of most meta-analyses. One of the most widely used tools to test for homogeneity is the Q statistic, despite evidence of its relatively poor performance. The three main arguments against relying on Q as a homogeneity test are relevant when applied to the standardized mean difference, g, as an index of effect size. These three weaknesses can be overcome by using different weighting schemes and transforming the values ​​to arrive at a closer approximation to the true distribution. We propose an alternative homogeneity test, to be used in meta-analyses with g values. It is based on a formulation of the random effects model as a mixture model and involves two normalizing transformations based on the true shape of the marginal distribution. A simulation study showed that our tests based on the transformations outperform Q, both in the probability of making a Type I error and in the resulting estimated power. Our method is also shown to be superior to Qñ, a variant of the classical Q with weights based on sample sizes, independent of the g values. The implications of the results are discussed, as well as how the logic of the proposed test can be extended to other effect size indices.

Zitieren

APA

Botella, J., Suero, M., & Durán, J. I. (2026). A HOMOGENEITY TEST FOR META-ANALYSIS WITH STANDARDIZED MEAN DIFFERENCES. Acción Psicológica. https://doi.org/10.5944/ap.46029

Vancouver

Botella J, Suero M, Durán JI. A HOMOGENEITY TEST FOR META-ANALYSIS WITH STANDARDIZED MEAN DIFFERENCES. Acción Psicológica. 2026. doi:10.5944/ap.46029

BibTeX
@article{Botella2026,
  author = {Botella, Juan and Suero, Manuel and Durán, Juan Ignacio},
  title = {A HOMOGENEITY TEST FOR META-ANALYSIS WITH STANDARDIZED MEAN DIFFERENCES},
  journal = {Acción Psicológica},
  year = {2026},
  doi = {10.5944/ap.46029}
}
Kostenlos starten →