Date: Wed, 1 Oct 1997 16:52:02 GMT
Reply-To: "K. A. Eder" <aeder@HSD.UVIC.CA>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: "K. A. Eder" <aeder@HSD.UVIC.CA>
Organization: Local Government Institute
Subject: Regression analysis with multiple completed data sets
I'm using the MVN macro from the SAS sample library to impute missing data.
The macro works well and I am able to generate multiple completed data sets
in one big dataset with an imputation number in the variable BY.
I am now performing regression analysis using the PROC MODEL. I would like
to generate a SINGLE set of results through some sort of averaging or
weighting procedure rather than the multiple results that a BY variable
outputs. I contacted SAS and they have no such programs to do this
weighting. They suggested that D Rubin's "Statistical Analysis with Missing
Data" might have details but I don't have access to that book right now (I
have to go through an interlibrary loan).
I would appreciate any help in this area as I am a relative newcomer to SAS
and statistical procedures with missing data.
Thanks in advance,
K. Anthony Eder
Local Government Institute
University of Victoria