Testing Categorical Moderators in Mixed-Effects Meta-analysis in the Presence of Heteroscedasticity. Issue 2 (3rd February 2020)

• Record Type: Journal Article
• Authors: Rubio-Aparicio, María; López-López, José Antonio; Viechtbauer, Wolfgang; Marín-Martínez, Fulgencio; Botella, Juan; Sánchez-Meca, Julio
• Is Part Of: Journal of experimental education. Volume 88:Issue 2(2020)
• Page Start: 288
• Page End: 310
• Publication Date: 2020-02-03
• Subjects: Meta-analysis -- mixed-effects model -- subgroup analyses -- residual between-studies variance; Education -- Experimental methods -- Periodicals; Educational psychology -- Periodicals; Education -- Periodicals; Psychology, Experimental -- Periodicals; 370.15
• DOI: 10.1080/00220973.2018.1561404
• Languages: English
• ISSNs: 0022-0973

Abstract

Abstract: Mixed-effects models can be used to examine the association between a categorical moderator and the magnitude of the effect size. Two approaches are available to estimate the residual between-studies variance, τ r e s 2 —namely, separate estimation within each category of the moderator versus pooled estimation across all categories. We examine, by means of a Monte Carlo simulation study, both approaches for τ r e s 2 estimation in combination with two methods, the Wald-type χ 2 and F tests, to test the statistical significance of the moderator. Results suggest that the F test using a pooled estimate of τ r e s 2 across categories is the best option in most conditions, although the F test using separate estimates of τ r e s 2 is preferable if the residual heterogeneity variances are heteroscedastic.