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: calculating alphas with imputed data


From   daniel klein <klein.daniel.81@googlemail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: calculating alphas with imputed data
Date   Thu, 9 Jun 2011 09:27:01 +0200

Justina,

I think you need to be a little bit more specific about what exactly
you mean by "alpha"? Most times I see "alpha" it is used to denote the
type I error in NHST. Since this quantity cannot be estimated, I
assume you mean something different. However, people use "alpha" to
denote the constant in a regression as in y = alpha + betaX +e. It is
also known in the context of reliability as Cronbach's alpha and there
are probably a lot of other situations where this greek letter is
used. Without knowing what it is you want to estimate it is not
possible to tell you how you would do it with multiply imputed data.

I assume you know the command to estimate "alpha" in non-imputed
datasets. I further assume you did check the supported -mi estimate-
estimation commands (-help mi estimate-) and did not find the command
you are looking for.

You may try to use -mi estimate ,cmdok : command- , where comand is
the command you would use to estimate "alpha" in non-imputed datasets.

If this does not work a good place to start might be here :
http://www.stata.com/support/faqs/stat/mi_combine.html

Combining point-estimates is relatively straight forward. Simply run
the estimation on each dataset and save the results. The mean of those
results is the MI-Estimator.

Best
Daniel
*
*   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