Skip to content
Prev 1709 / 5636 Next

[R-meta] rma.glmm and relative risk

Dear David

My feeling is that if the log-binomial model does not converge it is 
telling you something about the data-set, possibly that it is trying to 
estimate risk ratios which would, in turn, lead to impossible 
predictions on the response scale. I must confess though this is just a 
hunch and have have zero evidence to back it up.

Michael
On 05/09/2019 09:11, David Hajage wrote: