Skip to content

[R-meta] Back transforming Z to r after meta-analysis

2 messages · Hooren, Bas van (VBW), Wolfgang Viechtbauer

#
Dear,


Currently I am performing a meta-analysis of correlations. I first transform these correlations to Fishers'z using metafor, but for interpretation I would like to back-transform the Z to r. I have read a few papers by Adam Hafdahl where he shows that a direct back-transformation as suggested by e.g. Borenstein can lead to errors and would therefore like to use the method he proposed for back-transforming Z to r (e.g. https://bpspsychub.onlinelibrary.wiley.com/doi/full/10.1348/000711009X431914?).


I did not find any information on the implementation of any back-transformation methods and am wondering if any you know if there is an R-script that implements these transformations. Note: I have emailed the author (Adam) and he also was not aware of any implementation in R.



Groeten/Kind Regards,

Bas Van Hooren, PhD(c) | Nutrition and Movement Sciences | T: +31 6 51643957 | E: basvanhooren at hotmail.com | Twitter: @BasVanHooren
#
Dear Bas,

You are looking for transf.ztor.int():

https://wviechtb.github.io/metafor/reference/transf.html

See the example for how to use this (the example uses transf.exp.int but it's the same principle).

Best,
Wolfgang