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: OLS assumptions not met: transformation, gls, or glm as solutions?


From   "JVerkuilen (Gmail)" <jvverkuilen@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: OLS assumptions not met: transformation, gls, or glm as solutions?
Date   Tue, 18 Dec 2012 16:00:29 -0500

On Tue, Dec 18, 2012 at 1:24 PM, Laura R. <laura.roh@googlemail.com> wrote:
> Thank you very much for your support.
>
> I thought generalized linear models (this is what I meant with glm)
> support different distributions of the dependent variable y, not the
> residuals. My dependent variable and the residuals are both right
> skewed, so maybe glm with inverse gaussian would be good.

You're right, it's about the distribution of the dependent variable
and as Maarten said, only in the linear model (Gaussian GLM) is it the
case that the residuals line up so neatly with the dependent variable.
An IG or gamma GLM usually has a log link, so it's going to deskew
that way. One way to deal with the inferential uncertainty induced by
model choice is to try it a few different ways to see if your
substantive interpretation is sensitive to choice of specification.
But that can lead to a real rabbit hole of decisions, too.

I second Maarten's point about the fact that you're at the point that
your own substantive needs and knowledge are necessary.
*
*   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