Dear researcher,
I use a generalized linear mixed model (glmmTMB) to find the combined fixed and random effects of four factors. When I want to plot the results by boxplots with letters, I use (multcomp::cld) function, but the (emmean) is always negative; however, when I use the (lme) with the same parameter, I have positive (emmean). Is that normal ? and if yes, why have we negative emmeans values?
thank you in advance
Dear Dima,
in principle it is possible that emmeans take negative values, however given your puzzlement about this value, I assume that you deemed it incorrect.
With the information you provided, I doubt that someone can come up with a helpful reply. This is why you need to provide at least the following information to solve the mistery:
What are you trying to model? Are negative values possible? What link function did you use?
What functions from what packages were used? I assume you used the emmeans-function from the package with the same name but what options were included in the function call, especially for the mean function?
Please also provide the R output for the functions.
Bests,
Sven
I agree with Sven Greving ,
despite, one can try
https://cran.r-project.org/web/packages/emmeans/vignettes/transformations.html
summary(someEMM, ... , type = "response")
to express the response value on the actual scale, in case you
transformed it within the glmm
hth, Marian
- Badges
- Science topic