SPSS Statistics

 View Only
  • 1.  Hierarchical Cluster Analysis

    Posted Wed February 17, 2021 12:32 PM
    I've recently updated to SPSS 27 (from 25).  I previously ran a Hierarchical Cluster Analysis using between-groups linkage, Binary Squared Euclidean distance.  

    After updating to Spss 27, I cannot re-create the same output.  Running the exact same syntax does not give me the same cluster solution.  I am losing my mind thinking I must have done something differently, but I think it's something on the SPSS side that's changed.  I do not recall a 'present' and 'absent' box within the binary selection.  My data is binary, so that's the method I want to use, but what could have changed this?

    ------------------------------
    Lauriann Hebb
    ------------------------------

    #SPSSStatistics


  • 2.  RE: Hierarchical Cluster Analysis

    IBM Champion
    Posted Wed February 17, 2021 01:12 PM
    Based on the CSR, CLUSTER has not changed.  The Presence/Absence specification is  not new.
     
    "Different binary measures emphasize different aspects of the relationship between sets of binary values.
    However, all the measures are specified in the same way. Each measure has two optional integer-valued
    parameters, p (present) and np (not present).
    • If both parameters are specified, CLUSTER uses the value of the first as an indicator that a
    characteristic is present and the value of the second as an indicator that a characteristic is absent.
    CLUSTER skips all other values.
    • If only the first parameter is specified, CLUSTER uses that value to indicate presence and all other
    values to indicate absence.
    • If no parameters are specified, CLUSTER assumes that 1 indicates presence and 0 indicates absence."

    The agglomeration schedule table would be sensitive to case order, but the cluster assignment would not be except maybe for very small differences that could be affected by rounding.
    --





  • 3.  RE: Hierarchical Cluster Analysis

    Posted Wed February 17, 2021 03:55 PM
    I'm running the exact same syntax on the exact same data set.  The only thing that's changed is the update to SPSS.  Why would I be getting different results?

    ------------------------------
    Lauriann Hebb
    ------------------------------



  • 4.  RE: Hierarchical Cluster Analysis

    IBM Champion
    Posted Wed February 17, 2021 04:02 PM
    What things are different?  The cluster assignment?  The table?  What is your exact CLUSTER syntax?  Was V25 also the 64-bit version?  Are you on the same platform?

    --





  • 5.  RE: Hierarchical Cluster Analysis

    Posted Thu February 18, 2021 10:12 AM

    CLUSTER phone_user meeting_user event_user messaging1_user Admin_Y
    /METHOD BAVERAGE
    /MEASURE=BSEUCLID(1,0)
    /PRINT CLUSTER(3,7)
    /PLOT NONE
    /SAVE CLUSTER(3,7).

    This is the original cluster syntax I used.  I decided on the 4-cluster solution.  I saved my cluster membership.

    Using the same dataset and same syntax, I ran it again and got different cluster membership.

    Here is a crosstab of my original membership (named on the left in rows) compared to the current 4 cluster solution (1-4 columns).


    The new result is quite a bit different than the original.
    I was on a MacBook Pro for the first analysis, am on a new MacBook Pro currently (new computer = needed to re-install SPSS = upgraded to 27 from 25).



    ------------------------------
    Lauriann Hebb
    ------------------------------