Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: changing the order of the imputations while using ICE


From   "Francesco D'Amico" <francdamico@yahoo.it>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: changing the order of the imputations while using ICE
Date   Fri, 2 Dec 2011 11:54:55 +0000 (GMT)

Dear Martin, thanks a lot for your reply

I think you are completely right in saying that in general when using ice you should not worry about the order of imputations.


The reason why I was thinking to set the order of the imputations is to take advantage of the multiple imputation procedure whilst at the same time deciding the structure of the imputation one needs. In my case, let's suppose i have 5 variables (x1, x2, x3, x4 and x5). I know that x2 only depends on x1 and I don't want the other 3 variables to be included in the imputation. I can already set this by using the option eq() which allows me to specify which variables have to be inlcuded in any inputation equation. It annoys me though that this step cannot be set to be the first but happens in an already set position (say the third for instance), lacking the logical sequentiality I would need. 

The more logical solution could be to separate manually the two steps by using mi impute in the first step (let's say generating 5new imputations) and ice in the second step. The problem I would have in doing this is that I don't know what is the best way to include the 5 newly generated imputations in the second step based on ice.

Many thanks again

Francesco

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index