Statalist


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

Re: st: problem in ICE for multiple imputation


From   Craig Anderson <clanders@uci.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: problem in ICE for multiple imputation
Date   Thu, 12 Jul 2007 22:07:25 -0700

a
In a complicated imputation, I found it helpful to check each regression separately before running ICE. If it's not obvious from 2- way tabs where the problem is, you could even run ologit or mlogit with one independent variable and then add one variable at a time to see which one is preventing convergence.

Craig

On Jul 11, 2007, at 11:00 PM, Maarten buis wrote:


--- tang5@purdue.edu wrote:
A survey dataset has four variables q098 q099_2, q100 and 130 with a
heavy amount of missing values for each of them.
I ran the ice command and get the following message:

"I detected a problem with running uvis with command mlogit on
response q098 and covariates q099_2 q100 q130.

The offending command resembled:
uvis mlogit q098 q099_2 q100 q130 ,

With mlogit, try combining categories of q098, or if appropriate, use
ologit

unrecognized command
r(199);"

Then I use the option cmd(q098:ologit) since that variable could be
an ordinal variable. I got the following message:
"I detected a problem with running uvis with command ologit on
response q098 and covariates q099_2 q100 q130.

The offending command resembled:
uvis ologit q098 q099_2 q100 q130 ,

unrecognized command"
Try combining some of the categories, like I virtualy always do when I
use rep78 in my examples:
http://home.fsw.vu.nl/m.buis/stata/exampleFAQ.html#recode .

-mlogit- and -ologit- somtimes won't converge if one of categories is
very small. More seriously, the algorithm underlying -ice- (a Gibbs
sampler) won't converge if another variable or variables have all
missings for one value of q098. This is serious as -ice- will not and
cannot automatically check for convergance. I got the impression that
Patrick now checks for these problems and issues this warning message.

Hope this helps,
Maarten

-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------



___________________________________________________________
What kind of emailer are you? Find out today - get a free analysis of your email personality. Take the quiz at the Yahoo! Mail Championship.
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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   |   What's new   |   Site index