Hello,
I wonder to know what meta-analysis model must be used to compute a combined effect across subgroups within a study: fixed effect or random-effects? and why?
Thanks