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

st: Multiple imputation Using ICE

From   Sham Lal <[email protected]>
To   "[email protected]" <[email protected]>
Subject   st: Multiple imputation Using ICE
Date   Thu, 10 Jan 2008 17:20:04 +0000

Dear all, 
I'm currently in the process of analysing an imputed dataset but am having trouble performing correct regression models.
The current plugins used for ICE i think are up to date however I have specified them below just incase I have got the wrong ones. 
which ice
*! version 1.4.2 PR 03aug2007.
*! Based on _mvis2 version 1.0.2 PR 19jan2005.
*! v 1.0.0 PR 01Jun2001.
*! version 1.0.0 PR 20dec2004.
*! Based on artformatnos.ado v 1.0.0 PR 26Feb2004
*! version 1.0.0 PR 30may2006.

which mim
*! mim.ado v1.1.0 (JCG, JBC, PR)

which mvpatterns
*! 2.0.0  10jan2001   (STB-61: dm91)

once the dataset is imputed, i wish to regress 2 variables as follows
mim: regress active0_i chaq2_i 

however the following error occurs:
invalid 'Settings' 

Again if i run 
mim regress active0_i chaq2_i 

the following error occurs:
last estimates not found

I then tried using the micombine command:
micombine regress active2_i chaq2_i
another error "imputation identifier _j not found"

I have noticed that there is an _mj
so i create _j=_mj
and this seems to work with the following regression:
micombine regress  active2_i chaq2_i 

But i do not know if this is the correct way to analyse the dataset

additionally it does not work for 
xi: micombine mlogit active2_i subtype_i 
for which subtype_i was imputed as a categorical variable passively

any help would be appreciated, I have read through the UCLA guide on multiple imputation using ICE adn various other online resources but have yielded no results. And I keep hitting impassible brick walls


*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index