Need help with multiple imputation

[Chi-Chi Ojinnaka]

Good evening,

Please, I performed a multiple imputation using ice and saved it out as another data set. I tried to perform a logistic regression using mi:estimate on the new (imputed) data set but I get the error message "no imputations". Does that mean that I do not need to specify mi estimate even though the data is imputed?

Thanks.

[Alex Gamma]

Please post with your name and please read the FAQ. This is an elementary question whose answer you could easily find by taking a look at the documentation of the commands you used.

-mi: estimate- works with data imputed by -mi: impute-. You imputed your data using -ice-, which is a user-written command. Stata's -mi: estimate- does not know how to handle files produced by -ice-.

Alex

[Yulia Marchenko (StataCorp)]

You may also read about the relation between Stata's official mi command and the user-written ice command in the following FAQ:

http://www.stata.com/support/faqs/st...s-ice-and-mim/

[Richard Williams]

Here is a basic and a much more advanced intro to mi if you want it:

http://www3.nd.edu/~rwilliam/xsoc63993/l13.pdf

http://www.ssc.wisc.edu/sscc/pubs/stata_mi_intro.htm

I am not sure if there is any good reason to still use ice if you have one of the newer versions of Stata. But, as Yulia points out, mi can work with files created by ice.