Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • nested logit tree

    I am running a nested logit model of 2 levels. For level 1: I have 4 options.
    For the option 2 in level 1, I have only one choice that correpsonds to it in level 2.

    In other words, I have one option in level 2 and I do not have a lot of option like other nests.

    My question is : The results will be good? I think i will have problem in the inclusive value "dissmilarity parameters" for this option 2. As i dont have different nests in level 2.
    The dissimilarity parameter is equal to 1 for the option 2. What can I do. How can i interpret the result of the inclusive value? Can i make confidence in the results?

  • #2
    You face what is called a degenerated nest, although the nested logit could be run.
    There is a famous example circulating all around on the internet with a degenerated branch, talking about transport mode (with two nest air and ground, and a single option in the "air nest.) It cames from Green's (2003) book.
    You'll find here an application with Stata (at the middle of the page when talking about Example 19.18 on page 868).

    Charlie

    Comment

    Working...
    X