| |
 |
|
|
Science Forum Index » Statistics - Math Forum » GEE for a count variable and categorical predictor
Page 1 of 1
|
| Author |
Message |
| nyggus |
 Posted: Fri Jan 05, 2007 3:50 am |
|
|
|
Guest
|
Hi.
Suppose you have a following experiment:
Variable of interest: counts of an event (e.g., occurrence of a
pathogen) on three different parts of a plant.
Predictor (categorical): part of a plant (three levels).
Replication: plant.
Aim: to assess the difference in no. of counts of an event (no. of
pathogen organism) on different parts of plant.
How would you analyze the data? Of course, the observations on the same
unit (plant) are correlated. Generalized estimating equations are to
analyze such problems, but I haven't been able to find its application
for categorical predictors for count variable (e.g., with Poisson
distribution). The additional problem is that I want to study it in a
couple of environments, so it would be nice to include environment in
the model as well. Any ideas?
Thanks,
Nyggus |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Wed Oct 08, 2008 3:17 am
|
|