When generating an MDM document from the .sav file or accessing the case data with no metadata, you might find that the SPSS Statistics SAV DSC unexpectedly maps IBM SPSS Statistics categorical variables to MDM non-categorical variables, and vice versa. By default, the interprets an IBM SPSS Statistics numeric or short-string variable as categorical if it has one or more labeled values that are not defined as user-missing, or if the variable is part of a IBM SPSS Statistics multiple response set.

To get the SPSS Statistics SAV DSC to interpret IBM SPSS Statistics variables in the way that you want, you could edit the .sav file and add or remove value labels as appropriate, change whether the existing labeled values are defined as user-missing or not, or remove a variable from a multiple response set.

Alternatively, use one or more of following settings to control how the SPSS Statistics SAV DSC determines which variables are categorical:

(See Properties and settings used by the SPSS Statistics SAV DSC.)

These settings do not affect IBM SPSS Statistics variables that are part of a multiple response set, as the always interprets those variables as categorical.