2018
DOI: 10.1037/apl0000328
|View full text |Cite
|
Sign up to set email alerts
|

The heterogeneity problem in meta-analytic structural equation modeling (MASEM) revisited: A reply to Cheung.

Abstract: Yu, Downes, Carter, and O'Boyle (2016) introduce a new technique to incorporate effect size heterogeneity into meta-analytic structural equation modeling (MASEM) labeled full information meta-analytical structural equation modeling (FIMASEM). Cheung's (2018) commentary raises concerns about the viability of FIMASEM and provides its initial validation. In this reply, we briefly respond to those concerns noting how they relate to Yu et al.'s original conclusions, general MASEM practices, and operational decision… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
16
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
8

Relationship

0
8

Authors

Journals

citations
Cited by 19 publications
(16 citation statements)
references
References 27 publications
0
16
0
Order By: Relevance
“…We considered a number of multivariate tests using meta‐regression and meta‐analytic structural equation modeling. Although these tests are commonly employed in meta‐analyses, recent concerns over their validity when (a) the number of included studies is small (Viechtbauer, López‐López, Sánchez‐Meca, & Marín‐Martínez, 2015), (b) there are large amounts of true score variance (Yu, Downes, Carter, & O'Boyle, 2016), and (c) underlying assumptions are violated (e.g., Lv & Maeda, 2019; Yu, Downes, Carter, & O'Boyle, 2018) raise serious concerns over their appropriateness in many organizational behavior and human resource contexts.…”
Section: Resultsmentioning
confidence: 99%
“…We considered a number of multivariate tests using meta‐regression and meta‐analytic structural equation modeling. Although these tests are commonly employed in meta‐analyses, recent concerns over their validity when (a) the number of included studies is small (Viechtbauer, López‐López, Sánchez‐Meca, & Marín‐Martínez, 2015), (b) there are large amounts of true score variance (Yu, Downes, Carter, & O'Boyle, 2016), and (c) underlying assumptions are violated (e.g., Lv & Maeda, 2019; Yu, Downes, Carter, & O'Boyle, 2018) raise serious concerns over their appropriateness in many organizational behavior and human resource contexts.…”
Section: Resultsmentioning
confidence: 99%
“…This differs from the MASEM approach which is based on a correlation matrix of average effect sizes across different subpopulations. However, there was a statistical error (i.e., NPD: nonpositive definitive) when we used the FIMASEM approach to test our final model with cognitive and skill-based outcomes separated (see the discussion of Cheung [2018] and Yu et al [2018] about the NPD issue). To address the NPD issue, Cheung (2018) suggested replacing near positive definite matrices with NPD matrices in the two-stage FIMASEM (TS-FIMASEM) approach.…”
Section: Random-effects Fimasem Analytic Resultsmentioning
confidence: 99%
“…Second, Cheung (2018) shows that FIMASEM can be used to quantify the heterogeneity of the parameters with credible intervals. However, the validity of the indices of generalizability is questionable (see Yu, Downes, Carter, & O’Boyle, 2018, for the reply).…”
Section: Discussionmentioning
confidence: 99%