There are two solutions to your problem with ice. First, you can use
the match option. That will match your missing values with real
observed values. Second, you can impute with ologit if your range is
not too great. Sorry, I don't remember the range of the CESD exactly,
but I think you could do that too. Check out the help file and try
both methods separately.
Fred
On Sat, Oct 3, 2009 at 2:39 PM, Cathy L. Antonakos <[email protected]> wrote:
> Using -ice-, I created 5 imputed datasets. The imputed variable of key
> interest is a continuous variable. When I examine the values of the imputed
> variable, there are values outside the valid range, which I understand can
> happen when using -ice-.
>
> The key variable (depression, the CESD) is sometimes used as a dichotomy to
> determine how many cases lie at or above 16. An even higher cut-off may be
> used.
>
> I would like to create a dichotomous variable from the imputed continuous
> variable. Will I run into any trouble doing that? Should I examine the
> distribution, to see whether the percentage of cases at or above 16 is
> similar to that in the original data?
>
> Note that the CESD is a total score, based on a set of variables. However, I
> did not impute values for the individual variables. I imputed values for the
> CESD total score.
>
> I would be interested in a reference addressing this type of data
> manipulation using imputed data.
>
> Thank you for your help.
>
> Cathy Antonakos
> *
> * 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/
>
--
Fred Wolfe
National Data Bank for Rheumatic Diseases
Wichita, Kansas
NDB Office +1 316 263 2125 Ext 0
Research Office +1 316 686 9195
[email protected]
*
* 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/
