If you have access to SAS, Proc GLIMMIX generalizes Procs Genmod and 
Mixed, allowing hierarchical random effects and various link 
functions and error distributions. Like Proc Genmod, it is 
well-suited for binomial (logistic) or count data and allows a 
poisson (overdispersed or not) distribution. For a non-linear model, 
try Proc NLMixed.

Yetta

Reply via email to