home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!stanford.edu!bcm!convex!darwin.sura.net!paladin.american.edu!auvm!HARPO.DEV.UGA.EDU!GNELSON
- Message-ID: <SAS-L%92111209263065@UGA.CC.UGA.EDU>
- Newsgroups: bit.listserv.sas-l
- Date: Thu, 12 Nov 1992 09:26:31 -0500
- Reply-To: gnelson@HARPO.DEV.UGA.EDU
- Sender: "SAS(r) Discussion" <SAS-L@UGA.BITNET>
- From: gnelson@HARPO.DEV.UGA.EDU
- Subject: Conjoint Experts
- Comments: To: spssx-l@uga.cc.uga.edu, sas-l@uga.cc.uga.edu
- Lines: 51
-
- ----------------------------------------------------------------------
- CONTENT: Conjoint Analysis
- SUMMARY: Using a subset of an orthogonal array in conjoint
- REL/PLTF: SPSS 4.01/Mac, 4.0/MVS/CMS
- E-ADDR: gnelson@uga.cc.uga.edu
- NAME: Greg Nelson
- ADDRESS: UCNS - Client Services - Univ of Georgia
- PHONE: (706) 542-5359
- ----------------------------------------------------------------------
- This is being cross-posted. Disregard if uninterested in conjoint....
-
- Greetings! I am assisting a researcher here at the University of Georgia
- who would like to conduct a conjoint study which has 3 factors. It is a 6 x
- 5 x 2 design with 60 possible combinations. Obviously one cannot possibly
- digest 60 alternatives and discriminate but only a small amount of them.
-
- Question #1 - Is there any research which suggest the optimal number of
- alternatives that respondents can process?
-
- I then use the PLANCARDS procedure to obtain the orthogonal array. The
- programming looks like....
-
- ORTHOPLAN /FACTORS=
- BRAND 'Brand Name' (1 'A*' 2 'B*' 3 'C*' 4 'D*' 5 'E*' 6 'F*')
- PRICE 'Price' ('$7.99' '$9.99' '$11.99' '$13.99' '$15.99')
- FORMAT 'Formatted' (1 'Yes' 2 'No')
- /HOLDOUT=4.
-
- And produces (with the default seed) 49 experimental cards. That seems
- like a lot considering I started with only 60 possible. Furthermore, 49
- alternatives is also too many cards to digest.
-
- Question #2 - If I use a card sort procedure and have each respondent
- rank order a random subset (say 6 cards) I assume that SPSS will treat the
- others as missing. Is there a problem with that?
-
- OR ...
-
- since the researcher really wants all the levels of each factor retained in
- the study, can I randomly assign a subset of the levels and create several
- mini-studies using the entire orthogonal array for each mini-study? For
- example, if I use a 3 x 2 x 2 design I end up with only 9 cards. I could
- randomly generate all the possible subsets and have subjects card sort and
- then somehow combine the results of the "mini-studies:....?
-
- I appreciate hearing from anyone who has experience with designing and/or
- conducting a conjoint study.... TIA :-)
- Greg Nelson
- UCNS - Client Services
- The University of Georgia
- gnelson@uga.cc.uga.edu
-