Skip to content
Prev 5879 / 20628 Next

level 1 variance-covariance structure

Finally, I had to run the models using HLM, where I could get results 
(without iteration errors) for a model like this:

attit ~ age13 , data, random= ~ age13 | id, correlation = corAR1(form = 
  ~ wave | id))

I don't know why I could run this kind of models in HLM and I couldn't 
do it using R (lme). It would be good to know if it is a limitation of 
the R package or an over-parametrization of the model... or it is 
related to a specific characteristic of the dataset. I don't have any clue.

Thank you!
On 4/13/2011 3:29 AM, ONKELINX, Thierry wrote: