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

st: RE: Multiple imputation using ice error

From   "Johnson, Eric A. (Seattle)" <>
To   <>
Subject   st: RE: Multiple imputation using ice error
Date   Tue, 21 Oct 2008 14:12:36 -0700

You may have partial separation of your data with your logistic
regression command.  I believe logit will drop dummy variables that
perfectly predict a subset of the cases, and this is likely happening in
some (but not all) of your imputed datasets.  Try running
cross-tabulations between acr30 and the problematic variable, across
each imputed dataset--that should be able to detect whether my guess is
right or not!


-----Original Message-----
[] On Behalf Of Sham Lal
Sent: Tuesday, October 21, 2008 9:27 AM
Subject: st: Multiple imputation using ice error

Dear all, 
I have imputed a dataset using ICE, I have used it a few it along with
mim once before. However, with this dataset when I use mim I get the
following error message: 

Code : foreach x of varlist rs10498036 rs12477799 rs12995526 rs16853826
rs4672768 ///
rs4673990 rs13830 rs2295553 rs3810560 rs4815576 rs6037506 rs6051639
rs6051644 ///
rs6051655 rs6084305 rs6139035 rs10483080 rs12483377 rs1556329 rs17004785
rs2236475 rs2236479 ///
rs2274808 rs2838950 rs2838951 rs2838956 rs3753019 rs3788205 rs7499
rs9977268 rs11702425 {
xi: mim: logit acr30 i.`x', or

Error: covariates in analysis of imputed dataset 6 do not match those of
imputed dataset 1

It is able to perform some regressions but not all.

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index