Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: micombine and overall model fit statistics


From   "Rodrigo A. Alfaro" <raalfaroa@gmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: micombine and overall model fit statistics
Date   Mon, 21 May 2007 21:36:27 -0400

Claire,

As the 99% of the procedures Multiple Imputation (MI) is not magic. The command -micombine- collapses the results that you get from several simulated complete dataset and adjustes the standard errors to reconize that part of your data (the missing values) was "created" by some method (Bayesian, ICE, etc.). This is the way that MI "solves" the problem of missing values, basically you assign some values to replace the missing but you are ending with "more" datasets.

Usually, when the method works the R2's for each imputation are not too much different. I suggest you to report the average writing a small note about the dispersion of the all R2's that you get... maybe something like that in the table "R2 = 0.45" and in the note "Average R2 obtained from 5 imputed datasets. The min R2 obtained was 0.42 and the max 0.47". Assuming that the reviewers know what is MI, they will understand that the procedure worked for your case.





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