Dear all,
we have a choice problem with three alternatives, which are similar to three occupations among which a person can choose. Despite an alternative-specific constant we have only case-specific variables, such as personal charateristics. The IAA does not hold. Consequently, mlogit or mprobit is not suitable. But, alternatives such as nested logit (since we need at least two alternatives in each branch of the decision tree) or asmprobit and mixed logit (we do not have alternative specific variables) do not work either. What can we do? Thank you very much for any idea or hint. Gunnar
we have a choice problem with three alternatives, which are similar to three occupations among which a person can choose. Despite an alternative-specific constant we have only case-specific variables, such as personal charateristics. The IAA does not hold. Consequently, mlogit or mprobit is not suitable. But, alternatives such as nested logit (since we need at least two alternatives in each branch of the decision tree) or asmprobit and mixed logit (we do not have alternative specific variables) do not work either. What can we do? Thank you very much for any idea or hint. Gunnar
Comment