Page 1 of 1

Missing value in Longitudinal analysis

Posted: Sun May 30, 2010 7:56 pm
by zf8521
I'm doing the research on the longitudinal analysis for the health outcome. In the multinomail model, it can't contain any missing value. If I delete the missing valuse, the data structures will be unbalanced that can't run the model. Does anybody know, how to deal with the unbalanced data structures?