Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Multiple imputation and selection model

From   Maarten buis <>
Subject   Re: st: Multiple imputation and selection model
Date   Fri, 9 Mar 2007 10:10:22 +0000 (GMT)

--- Andréanne Léger <> wrote:
> I am working on a dataset with 2 problems:
> - missing observations for some countries that I assume are missing
> at random, and for which I use multiple imputation
> - some countries are not in the sample because no data are available,
> and this is not missing at random but rather linked to the level of
> institutional capacity/ development. I could use a Heckman selection
> model to control for this problem.
> Ideally, I would be able to address both problems at the same time -
> is it possible? Or should I solve the problems separately?

The terminology concerning missing data is a bit confusing. There is:

o Missing Completely At Random (MCAR); The probability of missingness
doesn't depend on anything.

o Missing At Random (MAR); The probability of missingness does not
depend on the unobserved value of the missing variable, but it can
depend on any of the other variables in your dataset, e.g.
institutional capacity / development.

o Not Missing At Random (NMAR); The probality of missingness depends on
the unobserved value of the missing variable itself, e.g. missing
values  in a variable "are you illiterate" from a written questionaire
are probably NMAR. 

The cases for which you wanted to use -heckman- sound like MAR to me
(unless the variables you want to impute are institutional capacity /
development). So, you can still use multiple imputation for these

Hope this helps,

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Inbox full of unwanted email? Get leading protection and 1GB storage with All New Yahoo! Mail.
*   For searches and help try:

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