Hello, I have a survey with missing data that I would like to impute, however included in this file are -8 values which indicate valid skips, which I do not want to include in the imputation or other descriptive statistics.
I can't seem to find a parameter to the imputation code (proc mi or proc mcmc) or function that would do the imputation and not take into account the -8 values. The file I wish to analyze is pretty big (50 x 15,000) so I can't just go in and edit the file.
I also looked at the proc Surveymeans, Surveylogic, Surveyreg, and Surveyfreq and I don't seen any function or parameter that would deal with valid skip variables. It's hard to believe this hasn't already been done. Do I have to read in each variable, strip out the -8, impute - find the means and other descriptive info and then put the -8's back in?? How have others dealt with this situation. Any help would be so much appreciated.