[R-sig-ME] Poisson GLMM: where is the deviance?

Ben Bolker bbolker at gmail.com
Sat Mar 17 22:03:55 CET 2012


Luciano La Sala <lucianolasala at ...> writes:

>
 
 [snip]

> Where "stake" (plots where the number of birds is measured) is as random
> effect, and "yobs" is the abundance of birds. I need an estimation of slope
> an intercept for the random term.
> 
> I specified my model as follows:
> 
> glmer(yobs~year + (year|stake), data=mibase,family=poisson)
> 
> and I get the following (it does not calculate the deviance)
> 
>   Generalized linear mixed model fit by the Laplace approximation
> 
  [snip]

> Q. What does this error mean and how can I solve it?
> 
> By the way, if I run the same model but only estimating the intercept of
> random effects as
> glmer(yobs~year + (1|stake), data=mibase,family=poisson) it seems to work
> properly.

  You aren't doing anything obviously silly, that I can see.

  Can we see a str() of your data?  I wonder if your 'year' variable
got turned into a factor by mistake.

  How many stakes do you have?  If it is fewer than 5-6 you are likely
to have numerical problems ...

  When in doubt plot your data ...




More information about the R-sig-mixed-models mailing list