[R] Meta-regression of categorical variables

[Kammy Tang]

Dear metafor users,


I am working on a meta-analysis of reliability and the correlation associations.


I need some help about conducting categorical moderators variables.


Questions 1: How to conduct the weighted ANOVAs assuming a mixed-effects model on the tranformed alpha coefficients/the tranformes correlation coefficients for the categorical moderator variables?


Questions 2: How to calculate the average reliability coefficients for each category of the moderator variable?


Take the reliablity generalization as an example. Here is an example of my code.


#meta-analysis using transformed alpha values (using the transformation suggested by Bonett, 2002)

res<-rma(measure="ABT", ai=ai, ni=size, mi=n_items, dat=data)
res
backtransf_res<-predict(res, transf=transf.iabt)
backtransf_res

#meta-regression of categorical variables
#take var1 as an example of categorical moderator variable
res_var1<-rma(measure="ABT", ai=ai, ni=size, mi=n_items, mods=~var1, dat=data)
res_var1
transf_res_var1<-predict(res_var1,newmods=c(1,2,3,4,5),transf=transf.iahw,digits=2)
transf_res_var1

NOTES: var1 consists of totally 5 levels/categories.

Is it appropriate to conduct like that as the example code showed?

Is that something wrong with the "newmods=c(1,2,3,4,5)" code?


To be honest, this is my first time to use metafor package and use R,  I definitely wish to find out the results of the meta-analysis.


Any help are appreciated!!

Thanks a lot!


Sincerely,

Jiaxin


	[[alternative HTML version deleted]]