[R] lmer - error message

ilai keren at math.montana.edu
Fri Feb 17 17:33:14 CET 2012


 I think my problem is that I can't
>> incorporate the 'lake' variable in a fixed-effect interaction because it is
>> only has one binary observation.  But I don't know what to do to be able to
>> fit this model.  Any help would be greatly appreciated!
>> -Sean
>
>  In principle you should be able to fit this model, but the error message
> is telling you that there are numeric problems -- it may just be that
> your data are a little too sparse in some direction.

Yes. Consider the collinearity between Age and Year, i.e. for a given
cohort (mos or all captured by "fishID" ?) they are essentially the
same variable with different units. So I would suspect the problem is
you are over fitting those.

A few suggestions:
>
>  * try centering Age, or re-introducing the intercept, to see if you
> can get something to work.
>  * You _might_ try the development version of lme4 (lme4Eigen, on r-forge)
>  * plot your data to see if you see anything odd about the data
>  * perhaps try making Year a fixed effect -- 4 levels is fairly small
> for a random effect
>  * Ask further questions on the r-sig-mixed-models mailing list.
>
>  Ben Bolker
>

Add one more to those: make sure your random effects are indeed
crossed. If they are nested (without knowing anything about your data,
just given their names that's a possibility), you could try nlme::lme.

Elai


> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.



More information about the R-help mailing list