[R-sig-ME] variance components models with zero estimates

Gregor Gorjanc gregor.gorjanc at bfro.uni-lj.si
Sun Jun 8 13:07:43 CEST 2008

David Airey <david.airey at ...> writes:

> When a variance components mixed model is run in Stata, if some of the  
> variance components are zero, the model may not converge, for rational  
> reasons according to the manual entry. However, when the same model is  
> run in SAS, the models with variance components that estimate to zero  

I remember that with lmer() you get a warning/message that says that
variance component for a particular effect is effectively zero in such cases and
the model converges!


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