[Limdep Nlogit List] Varying number of alternatives

[William Greene]

Dear M. Lagarde. 
Please see the chapter in the NLOGIT manual on specifying the
choice variables for choice models. There are specific ways
set up for exactly the situation you describe. You have to
provide a pair of additional variables in your LHS setup.
THe first tells how many choices the individual faces. The
second tells which. The ;CHOICES=list provides the list of
names in the universal choice set. Every individual will be
choosing among some or all of these.
/B. Greene
----- Original Message -----
From: "Mylene Lagarde" <Mylene.Lagarde at lshtm.ac.uk>
To: limdep at limdep.itls.usyd.edu.au
Sent: Saturday, August 15, 2009 1:33:51 PM GMT -05:00 US/Canada Eastern
Subject: [Limdep Nlogit List] Varying number of alternatives

I am trying to analyse a dataset where individuals have had to make a choice between a number of alternatives, but the number of alternatives varies from one individual to another - i.e. sometimes 4 of them (car / train / plane / bus), sometimes 3 of them, or even only 2.
I would like to estimate a mixed logit model, but don't know how to specify it that the choices individuals are faced with change for each choice set to the other. If I use the syntax "choices= car , train , plane , bus", then it doesn't work and I get the following error message:


  Error  1004: Number of observations not a multiple of # of alternatives.


How can I solve the problem?


Many thanks


Mylene


_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au