[R-meta] Principle component analysis VS multivariate meta-analysis
Viechtbauer, Wolfgang (SP)
wo||g@ng@v|echtb@uer @end|ng |rom m@@@tr|chtun|ver@|ty@n|
Wed Oct 16 18:30:26 CEST 2019
Just to clarify -- this is the r-sig-meta-analysis mailing list, not the metafor mailing list.
This aside, it's difficult to speculate about the motivation behind what somebody has done in this literature review without further details. But PCA is a variable reduction method and doesn't make any kind of model / distributional assumptions, so I am not even sure how to think about such a comparison (i.e., PCA vs multivariate meta-analysis).
From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces using r-project.org] On Behalf Of Wasim Iqbal (PGR)
Sent: Tuesday, 15 October, 2019 15:30
To: r-sig-meta-analysis using r-project.org
Subject: [R-meta] Principle component analysis VS multivariate meta-analysis
Dear Metafor members,
I was wondering if anyone could help unravel the reasoning behind carrying out a principle component analysis compared to a multivariate meta-analysis between outcomes.
For instance, a literature review collected data on enzyme kinetics from a range of published papers. They then analysed the relationships between the kinetic parameters using PCA. Surely, would a multivariate meta-analysis not be better since it will be taking into account any between study-heterogenity which would have otherwise been overlooked if a normal PCA was carried out?
Hope you understand what I mean.
More information about the R-sig-meta-analysis