sdm-help-list > Sample size
Showing 1-2 of 2 posts
Display:
Results per page:
Nov 2, 2022  11:11 PM | Pablo Arrona Cardoza - McGill
Sample size
Hello all! 

I have a general question about the sample size for the SDM meta-analysis. When I run the analysis, I get a message on the bias test table that reads: "WARNING: Metabias test conducted with < 10 studies". Because of this, we have a couple of questions:
Is this specific to the bias test? Or could I extrapolate it to the overall results of the meta-analysis? If this is the case, does it mean that 10 is the recommended number of studies to conduct the SDM analysis to have confidence in the results? I thought so, but we couldn't find that recommendation anywhere in the tutorial, manual, or any of the other SDM methods papers. Could this recommendation also be related to the total number of participants rather than the number of studies included?

Any comments on this will be highly appreciated it.
Nov 8, 2022  04:11 PM | Lydia Fortea - Instituto de Investigaciones Biom├ędicas August Pi i Sunyer (IDIBAPS)
RE: Sample size
Dear Pablo,

Although it is not mandatory, it is recommened to have a minimum of 10-12 studies for conducting a meta-analysis, depending also the effect size you want to achieve. Moreover, it can be reduced if you account with any statistical parametric map, as that will increase your statistical power. Moreover, it is also recommended that the included studies have a minimum of 10-12 subjects per group, for the same reasons. So far, there is no paper with the recommendation for conducting meta-analysis.

Best regards, 
Lydia