Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: MI: Multiple Imputation

From   Maarten buis <>
Subject   Re: st: MI: Multiple Imputation
Date   Mon, 20 Dec 2010 20:37:31 +0000 (GMT)

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen

--- On Mon, 20/12/10, Shikha Sinha wrote:
> I am using -mi for imputing the missing values. The
> variable is "edu" has five categories (1=primary,
> 2=secondary, and so on...). I used the following command:
> mi register imputed edu
> mi register regular age gender country
> set seed 23456
> mi impute mvn edu = i.gender age country, add (5)
> mi estimate, or: logistic Y X1 X2 edu
> After imputation, five new variables are generated, such as
> , _1_edu, _2_edu, _3_edu, .......
> Now, I have following questions:
> (a) Why some observations have negative values for 
> _1_edu, _2_edu, _3_edu, ........ Negative values for education
> doesnot make sense. How to restrict this so that there is no
> negative values.

You stated (by using -mi impute mvn- that education is normally
distributed, so Stata correctly concluded that negative values do 
make sense. If you don't like that you should not use the normal
distribution, but than it becomes hard to impute multiple 
variables simultatneously with the official Stata commands. 
Alternatively you can use -ice-, see -findit ice-.
> (b) How do we use the five new generated variable? Can I
> just run
> logistic Y X1 X2 _1_edu
> or
> logistic Y X1 X2 _2_edu

no, you have to use -mi estimate-.
> (c) Can I create a dummy for primary or secondary schooling
> from the imputed values _1_edu?

You should not do that, it tends to make things worse.

Hope this helps,


*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index