Hi Mike, and others,
I have a question on three-level meta-analysis and TSSEM. I'm working on a meta-analysis, in which I want to test the effects of three predictors on three outcomes, preferably using TSSEM. However, several studies provide multiple effect sizes for a single relation. To overcome this dependency, I would like to use three-level meta-analysis. However, to me as a beginner it looks like the 2 methods are not compatible. Therefore, to me, I have two options if I want to use MASEM: 1) Conduct analyses on bivariate relationships by fitting 3-level RE models, and use the correlation matrix as input, or 2) Aggregate effect sizes and conduct TSSEM. (Total effect sizes before aggregation = 260, after aggregation = 160)
What would you recommend?
Thanks in advance!
Printer-friendly version
Hi Rutger,
The preferred method is to conduct a three-level multivariate meta-analysis on the correlation matrices and fit the structural equation models on the average correlation matrix (https://books.google.com.sg/books?id=Pw7QBwAAQBAJ&pg=PA198&lpg=PA198&dq=6.4.3+Three-level+multivariate+meta-analysis&source=bl&ots=zCjT9lnFM1&sig=w4H6v0Sp1o43GdqQR6y3S2aVlAE&hl=en&sa=X&ved=0ahUKEwj-g62GpPHLAhWDBo4KHXSUA7sQ6AEIGzAA#v=onepage&q=6.4.3%20Three-level%20multivariate%20meta-analysis&f=false). However, it may be difficult to meet some of the assumptions required. Moreover, the metaSEM package has not implemented this yet though it can be done in OpenMx.
Both options sound reasonable. They have their own limitations. For the first approach, you may need to figure out how to use the average correlation matrix to fit SEM in the stage 2 analysis. For the second approach, we have to assume that the aggregated correlation matrices behave similarly to the observed correlation matrices. I would try the second approach first as it is easier.
Best,
Mike