Answers for "May someone help me to solve a problem with running Conjoint Analysis Syntax?"
https://developer.ibm.com/answers/questions/284272/may-someone-help-me-to-solve-a-problem-with-runnin.html
The latest answers for the question "May someone help me to solve a problem with running Conjoint Analysis Syntax?"Answer by BettinaW
https://developer.ibm.com/answers/answers/285074/view.html
Lulu17: This is was I was trying to explain on my first response. SPSS Conjoint use full profile approach only. If you say you have 12 cards on your orthogonal design plan, the persons who should evaluate them must be presented with all 12 cards (Stimuli) to rank/score them all and not only 8 cards out of the 12, as you need 12 variables with the scores for each card for each individual. If you do not have that, you cannot use SPSS Conjoint algorithm. It is not enough to have only 8 cards in the file, where the model needs at least 12, sorry.Thu, 30 Jun 2016 06:35:15 GMTBettinaWAnswer by BettinaW
https://developer.ibm.com/answers/answers/284527/view.html
@Lulu17: This is was I was trying to explain on my first response. SPSS Conjoint use full profile approach only. If you say you have 12 cards on your orthogonal design plan, the persons who should evaluate them must be presented with all 12 cards (Stimuli) to rank/score them all and not only 8 cards out of the 12, as you need 12 variables with the scores for each card for each individual. If you do not have that, you cannot use SPSS Conjoint algorithm. It is not enough to have only 8 cards in the file, where the model needs at least 12, sorry.Tue, 28 Jun 2016 06:35:55 GMTBettinaWAnswer by BettinaW
https://developer.ibm.com/answers/answers/284298/view.html
Thank you for your feedback :-)Mon, 27 Jun 2016 10:16:40 GMTBettinaWAnswer by BettinaW
https://developer.ibm.com/answers/answers/284283/view.html
You probably have different number of plancards compared to the variables with the ranking of the cards per each individual. Below is an example syntax that works.
I have 3 factors with each 3 levels plus 4 holdouts. If you print the cards you get 13 cards on the example.
In my data file I have 13 variables (PREF1 to PREF13) plus an ID variable. Per row for each individual ID I entered the card number they ranked on top (pref1) to the card they ranked on the bottom (Pref13) but each card needs to be ranked by each individual, this is the full profile approach we use in SPSS Conjoint. You cannot show only some cards to each individual, they need to rank them all, for the Conjoint algorithm that is available in SPSS.
For example ID 1 row looks like below in my file (see picture below). Each of the 13 card numbers are ranked.
I hope this helps
regards, Bettina
*Generate Orthogonal Design .
SET SEED 1.
ORTHOPLAN
/FACTORS=ausstatt 'Ausstattung' ( 1 'Klimaanlage' 2 'Colorverglasung' 3
'Schiebedach') farbe 'Farbe' ( 1 'rot' 2 'blau' 3 'weiß')
ps 'PS-Leistung' ( 75 '75 PS' 90 '90 PS' 115 '115 PS')
/OUTFILE='C:\TEMP\PKW.SAV'
/HOLDOUT 4
/MIXHOLD NO.
*Print Cards.
GET FILE='C:\TEMP\PKW.SAV'.
PLANCARDS
/FACTOR=ausstatt farbe ps
/FORMAT CARD
/TITLE ' Karte: CARD' .
*Conjoint-Analysis.
CONJOINT PLAN='C:\TEMP\PKW.SAV'
/DATA='C:\TEMP\PKW_DATA.SAV'
/SEQUENCE=PREF1 TO PREF13
/SUBJECT= ID
/FACTORS= AUSSTATT (DISCRETE) FARBE (DISCRETE) PS (LINEAR MORE)
/PRINT=ALL
/PLOT=ALL
/utility='C:\TEMP\pkw_utility.sav'.
![alt text][1]
![alt text][2]
[1]: /answers/storage/temp/8054-conjoint.png
[2]: /answers/storage/temp/8055-conjoint-syntax.pngMon, 27 Jun 2016 08:59:07 GMTBettinaW