Hi,
Thank you for reviewing my question, I have a question regarding PROC MI. I want to obtain EM estimates (mean and covariance) from proc mi. But the question is that is there a way to perform a regression in PROC MI, such as Y=b+aX (with dependent variables and independent variables), and obtain the EM estimates of Y and X?
I can only obtain EM estimates by using 'outem' with a regular 'var' statement, but it's not regression. I checked the manual, monotone and logistic statement can perform a regression, but that's not what I want, and I can't obtain EM estimates.
So is it feasible to obtain EM estimates by modeling a regression in proc mi?
Thank you!
