[Limdep Nlogit List] Variable choice sets

[Mikołaj Czajkowski]

On 2010-06-18 04:22, Yijing Lu wrote:
> Dear lister,
>
> I have encountered some questions when I estimated MNL. In my model, each
> individual is faced with 10 alternatives. However, the 10 alternatives which
> consist of one actually chosen alternative and 9 random selected
> alternatives from the universal choices sets. Therefore, different
> individuals have different choice sets. The difficulty for me is how to
> build the utility functions U(list), since the alternatives are not fixed.
>

Dear Yijing Lu,

If your utility functions are generic you may want to try to define 
explanatory variables using ;rhs= or you can specify them by U(*)= which 
will refer to all alternatives.

Best regards,

     Mikolaj Czajkowski