Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Discrete choice estimation

From   "Data Analytics Corp." <>
To   Stata Listserve <>
Subject   st: Discrete choice estimation
Date   Fri, 14 May 2010 10:52:22 -0400

Hi Stata,

I want to estimate a stated preference discrete choice model in Stata (Stata 11 on Windows). Let's say it's just a simple binary model with alternatives A and B. For a forced choice problem, I don't have a problem doing the estimation. I would use clogit with group(caseID) where caseID specifies each alternative for each choice set for each person. My problem is that a Neither is included. The original design matrix only has the two alternatives for the forced choice. The Neither was added in the questionnaire as it should be. Suppose my data consist of three independent variables (Color, Shape, Price), all categorical. It seems that the data for the first two choice sets, therefore, should be

RespondentID Set caseid choice Color Shape Price 1 1 1 0 red square low 1 1 1 1 blue round high 1 1 1 0 none none none 1 2 2 0 blue round high 1 2 2 0 blue round low 1 2 2 1 none none none

where the third and sixth rows reflects the Neither. Only rows 1,2,4,5 are from the design matrix where rows 1 and 2 are for alternative A and the second for B; similarly for 4 and 5. The last is added afterward to reflect Neither. Is this correct? The Stata command would be

        xi: clogit choice i.Color i.Shape i.Price, group(caseID)

How does Stata handle the estimation? Is the Neither alternative omitted (I suspect so)? If so, is the effect of Neither 0?




Walter R. Paczkowski, Ph.D.
Data Analytics Corp.
44 Hamilton Lane
Plainsboro, NJ 08536
(V) 609-936-8999
(F) 609-936-3733


*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index