Bounding of values in 2LevelImpute
Posted: Mon Dec 01, 2014 6:30 pm
I think the answer to this question is no, but I'll pose it anyway. Suppose I have an outcome variable (y) with missing data that has four possible values (0-3). Say the first two are for those who survive (x=0) and the last two for non-survivors (x=1). Therefore, imputations for missing data for x=1 must be from y=0 or 1, and for x=1 from y=2 or 3. Other imputation software like IVEware allow bounding of the imputations to account for these features in the data, but I can't find any way to do it in the 2LevelImpute template. Can it be done? Thx.