IBM Support

PM65582: PREDICTED CATEGORY CHECK BOX DISABLED IN GENLIN SAVE DIALOG FOR ORDINAL RESPONSE MODEL

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as fixed if next.

Error description

  • You are using the Generalized Linear Models procedure (GENLIN
    command) to build and fit an ordinal logistic model. You wish to
    save the predicted category for each case as a new variable in
    your active data set, along with the estimated probabilities for
    each category.
    From the Analyze menu, you choose Generalized Linear
    Models->Generalized Linear Models
    In the "Type of Model" dialog, you choose Ordinal logistic.
    In the "Response" dialog, paste jobcat into the Dependent
    Variable box
    In the Predictors dialog, you paste categorical predictors into
    the Factors box and scale level predictors into the Covariates
    box.
    In the Model dialog, you paste your predictors into the Effects
    box as Main Effects.
    In the Save dialog, you see that the check box for Predicted
    category is disabled.
    You click the box for Cumulative Probabilities and click the
    Paste button.
    In the syntax window, you add the keyword "predval" to the /SAVE
    subcommand. The  subcommand looks like this (but with your own
    variable names):
    
    GENLIN jobcat (ORDER=ASCENDING) BY gender (ORDER=ASCENDING) WITH
    minority educ
      /MODEL gender minority educ
     DISTRIBUTION=MULTINOMIAL LINK=CUMLOGIT
      /CRITERIA METHOD=FISHER(1) SCALE=1 COVB=MODEL
    MAXITERATIONS=100 MAXSTEPHALVING=5
        PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012
    ANALYSISTYPE=3(WALD) CILEVEL=95 CITYPE=WALD
        LIKELIHOOD=FULL
      /MISSING CLASSMISSING=EXCLUDE
      /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION
      /SAVE predval MEANPRED(:25) RESID.
    
    The command runs without error and the predicted category is
    indeed saved to the data set as a new variable. GENLIN was able
    to save the predicted category, so there appears to be no reason
    for disabling the "Predicted category" check box in the Save
    dialog of the Generalized Linear Models dialogs. A program
    defect was
    filed for this disablement. The current workaround is to build
    the command in the dialogs, click the Paste button to paste the
    syntax to a syntax window, add the PREDVAL keyword
    (capitalization not required) to the /SAVE subcommand as in the
    example above and run the GENLIN command.
    
    In the GENLIN dialogs, the Predicted category box is also
    disabled for Ordinal Probit models. It does not appear to be
    disabled for any model type other than ordinal response.
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED:                                              *
    * Users working with Generalized Linear Models                 *
    ****************************************************************
    * PROBLEM DESCRIPTION:                                         *
    * You are using the Generalized Linear Models procedure        *
    * (GENLIN                                                      *
    * command) to build and fit an ordinal logistic model. You     *
    * wish to                                                      *
    * save the predicted category for each case as a new variable  *
    * in                                                           *
    * your active data set, along with the estimated probabilities *
    * for                                                          *
    * each category.                                               *
    * From the Analyze menu, you choose Generalized Linear         *
    * Models->Generalized Linear Models                            *
    * In the "Type of Model" dialog, you choose Ordinal logistic.  *
    * In the "Response" dialog, paste jobcat into the Dependent    *
    * Variable box                                                 *
    * In the Predictors dialog, you paste categorical predictors   *
    * into                                                         *
    * the Factors box and scale level predictors into the          *
    * Covariates                                                   *
    * box.                                                         *
    * In the Model dialog, you paste your predictors into the      *
    * Effects                                                      *
    * box as Main Effects.                                         *
    * In the Save dialog, you see that the check box for Predicted *
    * category is disabled.                                        *
    * You click the box for Cumulative Probabilities and click the *
    * Paste button.                                                *
    * In the syntax window, you add the keyword "predval" to the   *
    * /SAVE                                                        *
    * subcommand. The  subcommand looks like this (but with your   *
    * own                                                          *
    * variable names):                                             *
    *                                                              *
    * GENLIN jobcat (ORDER=ASCENDING) BY gender (ORDER=ASCENDING)  *
    * WITH                                                         *
    * minority educ                                                *
    *   /MODEL gender minority educ                                *
    * DISTRIBUTION=MULTINOMIAL LINK=CUMLOGIT                       *
    *   /CRITERIA METHOD=FISHER(1) SCALE=1 COVB=MODEL              *
    * MAXITERATIONS=100 MAXSTEPHALVING=5                           *
    *     PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012               *
    * ANALYSISTYPE=3(WALD) CILEVEL=95 CITYPE=WALD                  *
    *     LIKELIHOOD=FULL                                          *
    *   /MISSING CLASSMISSING=EXCLUDE                              *
    *   /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION     *
    *   /SAVE predval MEANPRED(:25) RESID.                         *
    *                                                              *
    * The command runs without error and the predicted category is *
    * indeed saved to the data set as a new variable. GENLIN was   *
    * able                                                         *
    * to save the predicted category, so there appears to be no    *
    * reason                                                       *
    * for disabling the "Predicted category" check box in the Save *
    * dialog of the Generalized Linear Models dialogs. A program   *
    * defect was                                                   *
    * filed for this disablement. The current workaround is to     *
    * build                                                        *
    * the command in the dialogs, click the Paste button to paste  *
    * the                                                          *
    * syntax to a syntax window, add the PREDVAL keyword           *
    * (capitalization not required) to the /SAVE subcommand as in  *
    * the                                                          *
    * example above and run the GENLIN command.                    *
    *                                                              *
    * In the GENLIN dialogs, the Predicted category box is also    *
    * disabled for Ordinal Probit models. It does not appear to be *
    * disabled for any model type other than ordinal response.     *
    ****************************************************************
    * RECOMMENDATION:                                              *
    * This issue is resolved with IBM SPSS Statistics 21 release   *
    ****************************************************************
    

Problem conclusion

Temporary fix

Comments

APAR Information

  • APAR number

    PM65582

  • Reported component name

    SPSS STATISTICS

  • Reported component ID

    5725A54ST

  • Reported release

    J00

  • Status

    CLOSED FIN

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2012-05-25

  • Closed date

    2012-08-29

  • Last modified date

    2012-08-29

  • APAR is sysrouted FROM one or more of the following:

  • APAR is sysrouted TO one or more of the following:

Fix information

Applicable component levels

  • RL00 PSN

       UP

  • RL00 PSY

       UP

[{"Business Unit":{"code":"BU048","label":"IBM Software"},"Product":{"code":"SSCQ88K","label":"Statistics Desktop"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"19.0","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
29 August 2012