[R-meta] Partial residual plots and meta-analysis

[Cesar Terrer Moreno]

Dear All,

Using our beloved metafor and model selection, I got to the conclusion that my best model was of the form:

mymodel <- rma(es, var, data=dat , mods= ~ 1 + A + B*C)

I want to show in a figure (to be included in a paper) the relationship between es and the individual moderators, including the interaction, while taking into account the effect of the rest of the moderators. Although I don’t have experience in this matter, it seems that partial residual plots is what I am looking for. I have tried to use the package visreg for this purpose, but apparently it cannot handle a rma output:

visreg(mymodel)
Error in formula.default(fit) : invalid formula

Do you know how to visualise the individual effects of the moderators in a meta-regression? 
Thanks
César


	[[alternative HTML version deleted]]