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 on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: Box-Tidwell Test


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Box-Tidwell Test
Date   Tue, 26 Apr 2011 09:17:37 +0100

Although this looks almost hopelessly vague -- no explanation of what
commands you are using, no precise syntax, no precise output, no
explanation of "not working" -- an answer can nevertheless be
attempted. If your predictor is an indicator (a.k.a. dummy), no
one-to-one transformation can possibly alter its statistical
properties in key respects, as any mapping can be summarised by its
results,  0 -> c, 1 -> d, and is thus equivalent to a linear
re-expression.

Think of it geometrically: two points can be mapped using any
transformation t(.) you like, but they remain (at most) two points,
and thus the transformation  can be summarised by the straight line
joining them in the space (x, t(x)) !

Nick

On Tue, Apr 26, 2011 at 8:55 AM,
<lisenger@students.mail.uni-mannheim.de> wrote:

> For my thesis I am using logisitic regressions and I wanted to check my
> modells for linearity. One test that is commmon is the "Box-Tidwell"-Test,
> yet in my case it is not working. Both my dependent and independent
> variables are dummys and now I was wondering why the test is not working?
>
> I know there are other tests as well, however I really would like to know
> why the test is not working?
>
*
*   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