Skip to content

Negative and zero value on mixed models (moses selebatso)

1 message · Highland Statistics Ltd

#
Have a look at the pdf file that comes with the betareg package, and see 
whether the beta distribution (after converting it to the 0 - 1 range) 
is something that would work for you. If so, then use glmmTMB to fit a 
beta model with random effects.

https://cran.r-project.org/web/packages/betareg/vignettes/betareg.pdf

Such an approach assumes that you *cannot* sample a value larger than 1, 
or smaller than -1.
Alain