Announcement

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

  • Assumptions on Beta in the Gumbel distribution in Logit regressions in State?

    I am using the user build command mixlogit to estimate models of school choice behaviour in Stata. I am going to use my results in simulations I want to make sure I use the correct scale parameter when I add the random shock. Usually, in a logistic regression, the random shock to utility would be assumed to be standard Gumbel (with my=0 and beta=1). I do find indications in Arne's documentation that this is the case for mixlogit as well but it is not clearly stated. The same goes for the general documentation of various logit comands. Can anyone confirm that it is indeed the standard Gumbel that is used?

    Best
    Dany
Working...
X