Date:         Mon, 18 Aug 2008 13:21:13 -0700
Reply-To:     aihuawang@yahoo.com
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Ai Hua Wang <aihuawang@YAHOO.COM>
Subject:      proc mi issues
Content-Type: text/plain; charset=us-ascii

Dear List:

I am using proc MI to conduct multiple imputation. I understand that I need to specify the classification variables in the class statement. To impute the classification variable using the method available in SAS PROC MI it has to be monotone missing. But my data is not monotone missing. So under this circumstance how should I proceed? Could anybody shed some lights?

My dataset contains both continuous and classification variables. I understand I had better to use the MCMC statement to impute the continuous variables and the monotone statement for the classification variables. But I cannot specify the two statements at the same time. Does this mean I need to impute continuous variables and classification variables separately and combine them together at the end? Some research shows we need to inlcude as many variables as possible? Is this a conflict?


Thanks a lot,
Aihua