I want to estimate an exploded (or rank-ordered) logit model using the maximum likelihood procedure MAXLIK in GAUSS. Does anyone have written/seen one procedure that calculates the log likelihood for this type of model?
Aptech has an Application called Discrete Choice that contains a number of versions of the logit model, Adjacent Categories, Stereotype, Ordered, Multinomial, Conditional, Nested. It sounds like you might be interested in the Ordered Logit. It's also possible that the "exploded" logit model is a type of Multinomial Logit. Discrete Choice however doesn't contain a specifically name "exploded" logit model.