--- On Fri, 9/10/09, McDonald, Catherine wrote:
> I am working with a small data set with missing values for
> which I am trying to do multiple imputation with ICE. I have
> some data that are non-normal. Do I have to do
> transformations? I have seen a reference to Schafer
> (Analysis of incomplete Multivariate Data 1997) that says
> there is evidence that imputation methods can work even when
> the data are not normal. But I did not know if the ICE
> command would be one of these methods. Is there a ref that
> would support leaving non-normal data as is for multiple
> imputation with ICE? Are there steps that I need to take?
> What options for transformations would be possible?
That depends on the type of non-normality. The Schafer book
discusses a method that is more restrictive than -ice-. In
particular -ice- handels categorical, ordinal, and interval
censored variables more elegantly. So if that is your
non-normality problem than you definately do not want to
do any transformation. If you have a continuous variable
which is very skewed than you might want to transform that
first (e.g. take a log).
As to references:
Royston, P. 2004. Multiple imputation of missing values.
The Stata Journal 4(3): 227--241.
http://www.stata-journal.com/article.html?article=st0067
Royston, P. 2005a. Multiple imputation of missing values:
update. The Stata Journal 5(2): 188--201.
http://www.stata-journal.com/article.html?article=st0067_1
Royston, P. 2005b. Multiple imputation of missing values:
Update of ice. the Stata Journal 5(4): 527--536.
http://www.stata-journal.com/article.html?article=st0067_2
Royston, P. 2007. Multiple imputation of missing values:
further update of ice, with an emphasis on interval
censoring. The Stata Journal 7(4): 445--464.
http://www.stata-journal.com/article.html?article=st0067_3
Royston, P. 2009. Multiple imputation of missing values:
Further update of ice, with an emphasis on categorical
variables. The Stata Journal 9(3): 466-477.
http://www.stata-journal.com/article.html?article=st0067_4
Hope this helps,
Maarten
--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany
http://www.maartenbuis.nl
--------------------------
*
* 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/