ResNL posted on Wednesday, June 16, 2010 - 6:44 am
I have ten categorical outcomes that contain a lot of missing data. It is MAR since the missingness depends on endorsing certain questions (that are not in my datafile). How can I go about for handling this missing data? I've seen some imputation options in the manual but it appears that those options cannot be used with categorical outcomes. Thanks for your assistance.
You missingness will not be MAR if the data the missingness depends on are not part of the analysis. DATA IMPUTATION can be used with categorical outcomes. See the Version 6 user's guide which is on the website.