Q: Demographic variables don't enter into the model? Do you just incorporate them implicitly in the store?

A: I haven't introduced it at all at this point. And that's where we're going to come into this hierarchical structure. So I think it's a good point. And that is how will I model this? Well, one alternative is to say, well, I don't want to put any demographic variates and it's different across the stores, what I may want to start off with is to say, well, let's just have individual store models. I know nothing about the way that brands change across these suburbs. So that would give me this individual least square model, individual generalizing square models or these sure models for each of the stores. So the problem there is that even though I've got a lot of data, I don't have enough to really do this precisely for every store. So, for some stores I'm going to be able to say look, this would be a good pricing strategy but the standard error is going to be so large on these co-efficients that I'm not going to be able to come up precisely with a good statement that this is what I should do. Now, yes?

Back