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: xtmixed for non-parametric outcome


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: xtmixed for non-parametric outcome
Date   Wed, 19 Dec 2012 11:34:30 +0000

The nature of your outcome isn't clear beyond an assertion that it is
not normally distributed. Even that need not be a problem as the
marginal distribution of the response is often of secondary
importance.

In a simpler situation, suppose x is uniformly distributed and y is
generated by a + bx + normal error with mean 0 and some variance.

Is it better for the variance to be very small, so that the marginal
distribution of y is approximately uniform?

Or for it to be very large so that the marginal distribution is closer
to normal?

In your case, you might consider a transformation of the response.

On a related note, I don't think data are best described as
non-parametric; there are techniques so described, but in either case
the term is not informative. (Arguably, the best such techniques do
estimate some parameter such as probability(Y > X).)

Nick
On Wed, Dec 19, 2012 at 11:12 AM, Nikolaos Pandis <npandis@yahoo.com> wrote:

> I am trying to run random effects model for continuous data with 3 categorical predictors (xtmixed).
> My outcome is not normally distributed and neither are the residuals after fitting the model.
> Question: is anyone aware of any Stata commands for non-parametric correlated data.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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