Date: Mon, 19 Nov 2007 08:57:05 -0700
Reply-To: John Fiedler <johnfiedler@oreon.net>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: John Fiedler <johnfiedler@oreon.net>
Subject: Re: Is There a Way in SPSS/SAS/Stata to Do This Type of Conjoint
Analysis?
Content-Type: text/plain; format=flowed; charset="iso-8859-1";
reply-type=original
Hans,
What you are trying to do is a series of paired comparions with a rather
incomplete design. There is a great deal of literature on the subject going
back 80 years to Thurstone's 1927 paper, "The Law of Comparative Judgment".
In recent years, Jordan Louvier has written extensively on the subject.
I've found the best current implementation of the approach to be Sawtooth
Softwares Max-Diff procedure which on the front end produces a very
efficient design and on the back end utilized hierarchical Bayes Regression
to estimate importances at the individual level. The company'e technical
paper at http://www.sawtoothsoftware.com/download/techpap/maxdifftech.pdf is
a good review of approaches to your problem.
JOHN
----- Original Message -----
From: "Hans Zhu" <spsshanschen@gmail.com>
To: <SPSSX-L@LISTSERV.UGA.EDU>
Sent: Sunday, November 18, 2007 12:07 AM
Subject: Re: Is There a Way in SPSS/SAS/Stata to Do This Type of Conjoint
Analysis?
> Hi there,
>
>
> I have the following datasets including 10 questions which respondents
> have
> to answer following questions: "If UFA introduce a new store in your
> community, could you tell me which factor is most important?"
>
> 1. Staff's knowledge/Product Quality/Both;
>
> 2. Store Location/Offer only Lifestyle Product/Both;
>
> 3. Price/Staff's Knowledge/Both;
>
> 4. Product Quality/ Store Location/Both;
>
> 5. Price/ Offer only Lifestyle Product /Both;
>
> 6. Store Location/ Staff's Knowledge/Both;
>
> 7. Product Quality/ Offer only style product/Both;
>
> 8. Price/Staff's Knowledge/Both;
>
> 9. Staff's Knowledge/Offer only Lifestyle Product /Both;
>
> 10.Price/ Product Quality/Both;
> That is, each question compares two characteristics and the respondent has
> to state which of the two is more important or both. I now want to specify
> which of the factors is the most important overall? Which type of the
> *Conjoint
> Analysis is it? *, Is there a way in SPSS/SAS/Stata to do this type of
> Conjoint Analysis?
>
> Thanks for your help
>
>
>
> Hans Zhu
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@LISTSERV.UGA.EDU (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
| 
