GENMOD or NLMIXED or GLIMMIX

[SBS]

I am new to SAS and don' t quite understand if GENMOD, GLIMMIX. or NLMIXED is the right for below hypothesis testing - 

I want to isolate gender differences in preferences for a particular attribute of a product based on data available about their product purchases. I understand that I have to use a mixed Poisson or Negative binomial (in case of over dispersion) regression model as in my data - (1) I have same person buying multiple products (also, a given product might be bought by several people); and (2) the product attribute I am interested is a count variable.

In essence, I want to test the hypothesis: Women prefer products that are high onproductAttribute1 more than do men, even after controlling for price of the products".

I am currently running the following procedure in SAS:

proc genmod data=myData;
class GenderCosumer consumer_id;
model productAttribute1 =GenderCosumer + price  / type3 dist=negbin;
Repeated subject=consumer_id/sorted type=exch;

Is this right?

[StatDave · Accepted solution]

You could use either GENMOD, GEE, GLIMMIX, or NLMIXED to fit a model to a count response with repeated measurements on subjects. The choice largely depends on what type of model you want to fit for what intended purpose. GENMOD and the newer GEE fit a Generalized Estimating Equations model that is a population averaged model for population level inferences. GLIMMIX and NLMIXED can fit random effects models that are considered subject-specific models for inferences more at the subject level.

[SBS]

Thanks a lot for clarification StatDave! Could you share the link to the newer GEE procedure? Is this available on University Editions as well?