[R-sig-ME] Non-conformable arguments while predicting with glmmadmb_model

[Ben Bolker]

On 14-08-08 01:05 PM, Reuben Smit wrote:
> I am attempting to do an observed versus predicted assessment of a negative
> binomial glmm.
> 
> My first step was to simply use the predict( ) function with my glmmadmb
> object on the dataset used to generate the model:
> 
> glmm_dataset$prediction<-predict(glmmadmb_object, glmm_dataset,
> type="response")
> 
> But the error message results:
> Error in X %*% beta : non-conformable arguments
> 
> What has baffled me is that when I use the same glmmadmb_object to predict
> onto newdata, in the same way above, it runs the predictions fine, with no
> error message and credible results.
> 
> Does anybody have ideas about what is going on here?
> 

  Reproducible example please?

  Ben Bolker