[R-sig-DCM] Incorporating a "None" or constant alternative in a stated choice experiment

Data Analytics Corp. walt at dataanalyticscorp.com
Sat Jan 29 14:36:17 CET 2011


Good morning,

Does anyone have any experience incorporating a None alternative in a 
stated choice experiment.  I'm particularly interested in how it's 
handled in the estimation.  Since the design matrix can't include a None 
(there's nothing to design), how do people add it at estimation.  I 
merely add a constant row to each alternative list with the word "None" 
as the level and, if it's part of the experiment, 0 for price.  The SAS 
online manuals by Warren Kuhfeld show it being handled this way.  Any 
opinions/experience other ways to deal with it?

Thanks,

Walt

________________________

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



More information about the R-SIG-DCM mailing list