Skip to content

logistic regression for repeated measures

2 messages · Joel Bried, David Duffy

#
On Mon, 10 Aug 2009, Joel Bried wrote:

            
[SNIP]

Surely you will want to use all your data, given you don't have that many 
pairs.  Why not first look at GEE models, ignoring most of the 
complications, and see if there are any interesting effects.  Then 
decide if more complex random effects models are worthwhile.

David Duffy.