Skip to content
Prev 949 / 20628 Next

mixed-effects model specification question

On Tue, May 06, 2008 at 08:56:51PM -0400, Mark Kimpel wrote:
I'm a bit confused still - sorry - do you therefore have three batches
per rat, for a total of 36 batches (with two missing possibly)?  If so
then I still think that your model is over-parametrized, and that you
can eliminate the Tissue random effect, based on your model output
below (thanks for including it).
Can you specify exactly what useful information is being used in the
analysis that would be discarded if you took averages?
It's not clear to me that the conclusions will be different.  If not,
then the simpler analysis seems to be the better one.  I would suggest
that you compare them and see what happens.

Cheers,

Andrew