EXCLUDING higher order interactions in SPSS GLM repeated measures model

Discussion in 'SPSS' started by Ann Kronrod, Nov 15, 2011.

  1. Ann Kronrod

    Ann Kronrod Guest

    Hi,
    I am using SPSS19.
    I would like to run a GLM analysis with 2 within subject variables and
    3 between subject variables, EXCLUDING the 5-way interaction from the
    full model.
    Cannot find how to do that...
    Thank you
    Ann
     
    Ann Kronrod, Nov 15, 2011
    #1
    1. Advertisements

  2. Ann Kronrod

    Bruce Weaver Guest

    Please paste the syntax for your model. If you are not a syntax user,
    set up the model via the menus as you normally would, then exit via the
    PASTE button rather than the OK button. This will paste the command
    into a syntax window.
     
    Bruce Weaver, Nov 15, 2011
    #2
    1. Advertisements

  3. Ann Kronrod

    Bruce Weaver Guest

    I meant please *post* the syntax for your model. ;-)
     
    Bruce Weaver, Nov 15, 2011
    #3
  4. Ann Kronrod

    Ann Kronrod Guest

    Here you go:

    DATASET ACTIVATE DataSet5.
    GLM T_P_Th1 T_P_Pl1 T_PSY_Th1 T_PSY_Pl1 BY ST Tang order_tour
    /WSFACTOR=factor1 2 Polynomial f2 2 Polynomial
    /METHOD=SSTYPE(3)
    /PRINT=DESCRIPTIVE
    /CRITERIA=ALPHA(.05)
    /WSDESIGN=factor1 f2 factor1*f2
    /DESIGN=ST Tang order_tour ST*Tang ST*order_tour Tang*order_tour
    ST*Tang*order_tour.

    Thank you!
    Ann
     
    Ann Kronrod, Nov 16, 2011
    #4
  5. Ann Kronrod

    Bruce Weaver Guest

    It appears that GLM lets you have control over inclusion or exclusion of
    interaction terms that involve ONLY between-Ss effects, or ONLY
    within-Ss effects, but NOT interactions that involve both between- and
    within-Ss effects.

    What you could do is restructure the data from WIDE to LONG (via
    VARSTOCASES), and then perform your analysis using MIXED--it will allow
    you to specify exactly what terms you want in the model. There are some
    examples in the Help showing how to perform repeated measures ANOVA via
    MIXED.

    HTH.
     
    Bruce Weaver, Nov 16, 2011
    #5
  6. Ann Kronrod

    Ann Kronrod Guest

    Thanks a lot, this is actually an interesting suggestion!
    I will try it out, definitely. Thank you!
     
    Ann Kronrod, Nov 16, 2011
    #6
  7. Ann Kronrod

    Bruce Weaver Guest

    You're welcome. And of course, as the Help file examples will show,
    you'll need an ID variable, given that you'll have multiple rows of data
    per subject (one row for each combination of the variables with repeated
    measures).
     
    Bruce Weaver, Nov 16, 2011
    #7
  8. Ann Kronrod

    Ann Kronrod Guest

    OK, will do that. Thanks a lot for the tip!
     
    Ann Kronrod, Nov 17, 2011
    #8
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.