Finding cases included in multinomial regression results

[Jamison, Cheryl S]

Hi List—

    Is it possible to run a syntax for multinomial regression which would either list the cases (by caseid) that ARE included or are NOT included (whichever is easier).  I have a datset with 712 cases, but the analysis only uses 454  cases.  I realize that cases are deleted if they have missing values, but there are 13 variables in the analysis (322 total variables in the dataset), so I’d like to avoid going through the dataset line by line to identify the missing cases, if possible.

      Thanks so much—

          Cher

[ViAnn Beadle]

Use the COUNT transformation with keyword missing to count the number of missing values in the list of specified variables. In this case that list will be all the variables used in your regression. You can then use the result variable to  filter or select cases and list them using the list command. Or, if you already have a case identifier variable in the file, reverse sort on the result variable and just look at the cases in the data editor.

 

COUNT  misvars= blah blah blah(MISSING).

FILTER BY misvars.

LIST VARIABLES = blah blah.

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jamison, Cheryl S
Sent: Wednesday, February 25, 2009 7:00 AM
To: [hidden email]
Subject: Finding cases included in multinomial regression results

 

Hi List—

    Is it possible to run a syntax for multinomial regression which would either list the cases (by caseid) that ARE included or are NOT included (whichever is easier).  I have a datset with 712 cases, but the analysis only uses 454  cases.  I realize that cases are deleted if they have missing values, but there are 13 variables in the analysis (322 total variables in the dataset), so I’d like to avoid going through the dataset line by line to identify the missing cases, if possible.

      Thanks so much—

          Cher