SPSS Statistics

 View Only
Back to discussions
Expand all | Collapse all

SPSS Conjoint does not compute an standard error for 4x3 Orthogonal Design

  • 1.  SPSS Conjoint does not compute an standard error for 4x3 Orthogonal Design

    Posted Fri September 06, 2024 10:46 AM
    Reply

    Dear IBM-Team, 

    For a conjoint study, I created an ORTHOPLAN 4 characters and 3 levels each. I received 9 cards and had the stumuli evaluated. I carried out the evaluation using the CONJOINT syntax according to the instructions. In the output it does compute the the part worths of the levels but unfortunately no standard error (see picture) and the following info message in the syntax window: "There are not enough degrees of freedom to calculate the standard error. "

    Any other combination i tried in an ORTHOPLAN with more or less characters and levels works fine with the right number of cards.  There are no errors or blanks in the dataset either. 

    When set the Minimum number of cards in the ORTHOPLAN-Tool to e.g. 12, it automatically jumps to 16 cards. When I then put Dummy-Ratings on top of my 9 original ratings to match the 16 cards, SPSS does compute the standard error all of a sudden. When I just try to put 7 holdout-cards into the Orthoplan it will not compute a standard error. 

    With 9 Cards and the same data I do get a standard error in R (conjoint package).

    what could be the problem in SPSS? 

    Looking forward to any support.



    ------------------------------
    Lutz Fischer
    ------------------------------