st: Formatting variables in ice

[fjgallo@mac.com]

Dear All,

I am new to Stata, and using version 10.0. I plan to use the ice  
procedure for the missing values of variables, which are MAR. All  
values are means of variables scaled on a continuous metric. There are  
no missing values of the criterion variable. The predictor variables  
have some missing values. However, the ice dry run suggests for  
example, mlogit for a variable that has only four values, but those  
values are means derived from ratings along a continuous scale. So  
should I change the mlogit to regress? I greatly appreciate any  
thoughts on building the appropriate imputation model using ice. Thank  
you.
Best regards,
Frank

 
 
*
*   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/