There is currently a bug in the cross-validation code for models that incorporate random effects (in either the occurrence or detection parts of the model), which can result in deviance metrics that may be inaccurate depending on the number of data points within each level of the random effect. I am currently fixing this and it should be updated on CRAN by mid-late March. For now, the waicOcc function can be used for model comparison when comparing models that have random effects.
There is currently a bug in the cross-validation code for models that incorporate random effects (in either the occurrence or detection parts of the model), which can result in deviance metrics that may be inaccurate depending on the number of data points within each level of the random effect. I am currently fixing this and it should be updated on CRAN by mid-late March. For now, the
waicOcc
function can be used for model comparison when comparing models that have random effects.