I want to do a multiple imputation of missing data using SPSS. I have a nominal (categorical) variable X with missing values codified as '9' and not applicable values codified as '8'.
How can I constrain imputed values to X, for not imputing '8' (not applicable) to a '9' (missing value)?
Since this variable X is categorical it is not possible in SPSS to use 'Define Constraints' for that purpose. Is this possible using other application for multiple imputation (e.g. R)?