How to use semEff with glmmTMB? about semeff

[murphymv]

Hi Sina,

Apologies for delayed response. Currently the package doesn't officially support glmmTMB models. The list of supported models is provided in the README. Supporting new model types is based on user demand and my time, but in some cases might not be too difficult. I don't currently have much time to do that though, so maybe you could use an alternative supported model type (e.g. glmer) in the meantime? Or you are always welcome to contribute code that provides support.

On the other question, yes you can use mixed and non-mixed models in the same SEM, if it makes sense to do so. The bootstrapping will however treat all models as mixed, as it will be assumed that the data is hieararchical and will resample based on the (highest-level) group (random effect). So it will resample groups instead of rows (that's if you use the default sampling type of 'nonparametric'). Is there a reason why you are not using mixed models for the lm() models? Usually you should do so if the data is hiearchical, unless the random effect variance is close to zero.

Cheers,
Mark