Scott,
I cannot address your questions about SOLAS at this time.  It has been a 
long time since I tried to use it.  However, there is a problem using the 
minimum and maximum options in PROC MI in SAS version 8.2.  If you use 
these features, you should examine your imputed data sets VERY CLOSELY.

I initially used these options in my dissertation analyses last summer and 
found that SAS was actually inserting values into cells where valid data 
existed in my input data set.  I contacted SAS about this and their 
solution was to remove the minimum and maximum values and then recode any 
values outside of the acceptable range to the nearest extreme that was 
within the acceptable range.

I am not saying that you should not use PROC MI.  The current bugs are 
being fixed for the next release and the current work-arounds, while a 
little inconvenient do appear to work well.  If I need to do imputation in 
the future I will likely use PROC MI.

Hope this is helpful.
Mike.

At 4/14/2002 09:43 AM, Scott Millis wrote:
>Dr Little:
>Thank you for the update regarding SOLAS. Perhaps I'm misinformed, but my 
>understanding is:
>
>(1) SOLAS does not offer a MCMC imputation method;
>(2) One cannot set minimum and maximum values for variables in the 
>imputation process (as can be done in PROC MI);
>(3) SOLAS cannot perform multiple imputation for logistic regression or 
>survival models (as can be done with Harrell's transcan).
>
>Clarification of these questions/impressions would be appreciated.
>
>SR Millis
>

E. Michael Bohlig, Ph.D.
1125 W. Meseto Ave.
Mesa, AZ  85210
480-775-1557  (home)
480-390-5434 (cell)

Reply via email to