[R] Problem with specifying variance-covariance matrix for random effects (nlme package)

[Uwe Ligges]

On 14.02.2010 21:20, Benjamin Cheah wrote:
> Hi all,
>
> I've been struggling with trying to specify a diagnoal matrix for linear mixed effects model. I think I've got nearly everything correct, except the following message appears:
>
> In lme.formula(fixed = fwave ~ sex + sexXbulbar + visit + age +  :
>    Fewer observations than random effects in all level 1 groups
>
> Not sure if i've provided enough details, but I'm basically trying to perform a mixed effects model analysis, controlling for several covariates. Incorporating random effects for the intercept and slope.
>
> What does the second line, 'Fewer observations than random effects in all level 1 groups' mean?

That your data is lacking degrees of freedom? In short: You need at 
least one observation for each parameter estimate to be able to estimate 
any coefficients, and you need much more in order to get sensible 
results. Hence you need either more data or need to estimate less 
parameters.

Best,
Uwe Ligges



>
> Many thanks in advance,
>
> Ben Cheah
>
>
>
>        __________________________________________________________________________________
> [[elided Yahoo spam]]
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.