I need to calculate Fail-Safe N of my meta-analysed studies. I am having difficulty in understanding Orwin's method. Can anyone please elaborate with the help of an example?
Orwin's fail-safe N is based on the concept of the "trivial" effect size and calculates the number of missing studies to bring the true effect size under that "trivial" effect size; however, what would be decided as "trivial" effect size is based on the specific scenario one may encounter. For example, in case of a common disease, a small reduction in mortality is clinically significant; therefore, the "trivial" effect size would be as small as you deduce from what is already known or what you wish to check. "Comprehensive meta-analysis ver 3.0" (https://www.meta-analysis.com/) is a valuable software that offers a free trial, through which you may well be familiarized in practice with the method. However, as Orwin's fail-safe N is unweighted, its use and interpretation is rather risky. Perhaps you may have a look at some weighted fail-safe N tests, like that one: Article The file-drawer Problem Revisited: A General Weighted Method...
I'm not sure Anas Iftikhar, but it sounds like you are trying to provide some assessments for publication bias? In that case, you could consider using some alternative tests given some of the potential limitations with using the fail-safe n method. Quantitative tests, like the Egger regression test and the Begg and Mazumdar rank correlation test, are popular methods for measuring funnel plot asymmetry. Where there is asymmetry, you can also consider the trim-and-fill method.