Parameters estimated in Proc MI

[deleted_user]

Hi there,

I'm a pretty new SAS user and I'm stumped on the output from PROC MI. I'm using the Regression method for Monotone data to impute missing values for one variable in my dataset, using 6 other continuous variables as predictors.

In the 9.1 User's Guide, it states that with the DETAILS option (i.e., monotone reg (Y = X 1 X2....X6/ details) the parameters estimated from the observed data and the parameters used in each imputation are displayed under the heading "Regression Models for Monotone Method".

My code runs fine, error free, and produces imputed datasets that seem reasonable; however the parameters estimated for each effect (i.e., independent variable) make no sense. I'm guessing that they are somehow standardized, but the manual makes no mention of this...If they are standardized, can someone please tell me what formula the program is using to do the standardization?


Cheers and thanks in advance,
Jenn

[deleted_user]

Plz try different K values in options ( DETAILS and K=).

[CGM]

Hi, Jenn.
I have the same question that you had. Did you find out anything?