sdm-help-list > About the permutation times
Showing 1-2 of 2 posts
Jun 20, 2017 07:06 AM | Zhiqiang Sha
About the permutation times
Hi,
I have two questions about SDM-based meta-analysis.
1) When we calculated the effect size of each study, the default value of Monte Carlo randomization is one, however, in a real meta-analysis several randomization are recommended. So, how many times do you suggest?
2) When we assess the statistical significance, the software have automatically performed the permutation test. So i do not know how many times of permutation tests?
Best wishes,
Zhiqiang Sha
I have two questions about SDM-based meta-analysis.
1) When we calculated the effect size of each study, the default value of Monte Carlo randomization is one, however, in a real meta-analysis several randomization are recommended. So, how many times do you suggest?
2) When we assess the statistical significance, the software have automatically performed the permutation test. So i do not know how many times of permutation tests?
Best wishes,
Zhiqiang Sha
Jun 20, 2017 09:06 AM | Joaquim Radua
RE: About the permutation times
Dear Zhiqiang
the default number of randomizations is small to allow the user conduct a quick check of the data (e.g. to see whether the maximums of the recreated studies are approximately in the locations of the maximums reported in the papers). However, a real meta-analysis should have more randomizations; we found that p-values were rather stable with only 20 (http://dx.doi.org/10.1016/j.eurpsy.2011....), but I would suggest some more, e.g. 50.
With the best wishes,
Joaquim
the default number of randomizations is small to allow the user conduct a quick check of the data (e.g. to see whether the maximums of the recreated studies are approximately in the locations of the maximums reported in the papers). However, a real meta-analysis should have more randomizations; we found that p-values were rather stable with only 20 (http://dx.doi.org/10.1016/j.eurpsy.2011....), but I would suggest some more, e.g. 50.
With the best wishes,
Joaquim