[R-sig-ME] degrees of freedom in mixed model

[S Ellison]

Just for the record, I just noticed that there was a rather crucial 'not' missing from my initial post, just before 'interpretable' in 
> So F is interpretable in the same way as it would be in classical anova.

But others have covered it more than adequately, and it's good to be reminded that using K-R is indeed not often a serious concern in LMMs despite the sometimes heated debate.

Me, I just wish I could find a good, simple answer for fixed effects confidence intervals with crossed random effects in small experiments ...

Steve E




*******************************************************************
This email and any attachments are confidential. Any use...{{dropped:8}}