Skip to content
Prev 6256 / 20628 Next

much different results for random effect vs simple lm.

On Mon, 20 Jun 2011, Brent Pedersen wrote:

            
You need to look at just a few probes in detail.  Given you have such a 
small sample size (and how many concordant pairs?), you might expect a bit 
of shifting about.  The other model you should check (in a subset) is your 
first model fitted to all 98 pairs, using your conservative degrees of 
freedom from model 2 (this would be pretty similar to a GEE, AIUI).

Cheers, David Duffy.