Re: st: ICE conditional option and categorical variables

[A Loumiotis <antonis.loumiotis@gmail.com>]

and in the second specification of the ice command what option do I
use to set catvar1a=0 if catvar1b is imputed to 1 and vice versa?


On Thu, Nov 18, 2010 at 4:39 PM, A Loumiotis
<antonis.loumiotis@gmail.com> wrote:
> Hi,
> I have a question in ICE. Suppose I have five continuous variables
> (var1-var5) and a categorical variable (call it catvar1 with 3)  all
> with some observations missing.  Suppose also that catvar1 defines
> conditions for other variables to be answered ie, var1: catvar1==1,
> var2: catvar1==2, var3: catvar1==4.
>
> In the following ice command is the conditional option correct?
>  ice var1-var5 m.catvar1, ... cond(var1: catvar1==1, var2:
> catvar1==2, var3: catvar1==4)
>
> Or should I transform it in terms of the corresponding dummy variables
> of catvar1 and adjust the mainvarlist accordingly?
>  ice var1-var5 catvar1a catvar1b, ... cond(var1: catvar1a==1, var2:
> catvar1b==1, var3: catvar1a==0 & catvar1b==0)
> where catvar1a=1 if catvar1==1
> and    catvar1b=1 if catvar1==2
>
> It seems to me that the latter is the case but I'm not sure.  For sure
> if the latter is the case then matters are complicated when someone
> has many categorical variables as basis for conditions.
>
> Thanks,
> Antonis Loumiotis
> *
> *   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/
>

*
*   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/