Currently, the development has been based on mean-differences ... need to check whether all models make sense for correlations (and ORs). Also, with correlations, r-to-z transformation should be used prior to meta-analysis, but then back-transformed for reporting - should we add a transformation function between MA and reporting? Maybe also relevant for other ES types?
Currently, the development has been based on mean-differences ... need to check whether all models make sense for correlations (and ORs). Also, with correlations, r-to-z transformation should be used prior to meta-analysis, but then back-transformed for reporting - should we add a transformation function between MA and reporting? Maybe also relevant for other ES types?