Re: Re: st: heckprob using multiple imputation

[Klaus Nowotny <klaus.nowotny@sbg.ac.at>]

Dear Marcus,

thanks, I was alredy aware of the posting. I am currently more inclined 
to use the second model and impute both y and income (either by mvn or 
ice, since y is binary) to preserve the common patterns in the data, but 
to drop the imputed values for y before estimating the outcome equation. 
In any case, at least in my data it does not seem to make any 
statistically significant difference whether y is imputed along with 
income or not, but still, input from an expert on the topic would be 
highly appreciated.

Klaus
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/