Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: R: imputation of missing dates


From   maarten buis <maartenlbuis@googlemail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: R: imputation of missing dates
Date   Fri, 6 May 2011 15:14:48 +0200

The fact that the probability of participating depends on education
and age is not a problem as long as you include it in your imputation
model. This is what (the rather confusing) term Missing At Random
means, missingness is random _after_ controlling for observed
variables. However, it sounds to me like you want something that is
logically impossible: to find a date at which an exam took place when
no exam took place. Remember that the whole machinery of imputation is
based on the idea that the value exists but was not observed, in your
case the value did not exist in the first place.

You might try matching. That will try to find for each examer a
non-examer that is as similar as possible on a set of observed
explanatory variables. The real problem is that people who choose not
to participate tend to do that for a reason, e.g. they did not choose
to participate because they already knew they would perform badly. If
such reasons are also strong predictors of the dependent variable
(which is often the case) and they are not measured (which is also
often case), then matching will not help you.

Hope this is not too depressing,
Maarten

On Fri, May 6, 2011 at 2:27 PM, Yulia Blomstedt <sungurik@gmail.com> wrote:
> Thank you Carlo.
> I've been going through the literature but have not yet been able to
> come up with  mi syntax that would work for me.
>
> Answering your questions:
> 1) The examination was voluntary. 60% of the invited individuals have
> participated. Others have not due to various reasons
> 2) Our earlier study showed that non-participants differ in their
> education level (higher) and age (younger) from the participants, but
> not in other characteristics.
> This of course raises a concern of whether mi is applicable at all,
> since its assumption is that the missing data are missing at random.
>
> If mi is not applicable, what other method could I use to create the
> EXAMDATE for non-participants?
>
> /Yulia
> *
> *   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/
>



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


http://www.maartenbuis.nl
--------------------------

*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index