Aggregate discrete choice models

[pmsas]

I'm looking for any information or sample programs using SAS to estimate a Berry-Pakes-Levinsohn type aggregate discrete choice logit model? These are random coefficient logit models using aggregate share data. 

Thanks.

Pat

[Ksharp]

I am not sure . But check PROC BCHOICE .

[pmsas]

Thanks for the suggestion. Unfortunately, BCHOICE is for individual rather than aggregate data, i.e. the response variable is a 0 or 1. 

Pat

[ballardw]

---

@pmsas wrote:

Thanks for the suggestion. Unfortunately, BCHOICE is for individual rather than aggregate data, i.e. the response variable is a 0 or 1. 

Pat

---

Do you have access to the data that was used to make the aggregate? That might be appropriate for BCHOICE.

You may want to provide a bit more description of your data such as is your aggregate a count or some other summary statistic(s).