Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: RE: Testing for interaction within linear regression


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: RE: Testing for interaction within linear regression
Date   Tue, 2 Sep 2008 12:33:01 +0100

All true, but a simpler answer is: -search interaction- (to find out for
yourself) and  -help xi-. 

(Also: please edit emails for previous irrelevant material.) 

Nick 
n.j.cox@durham.ac.uk 

Martin Weiss

I do not really know what to make of the normality issue; there is no
requirement on the covariates being normal, AFAIK. If you want to test
the
validity of your interaction, -generate- the product of the two and add
it
to your covariates. Then see what kind of an effect (sign, p-value)
comes
out of it. Be warned, though: We had a few posts recently about -nl- and
how
it is able to see that you used the same covariate twice and adjust the
resulting marginal effect for -mfx- while -regress- would not know that
you
threw in the covariate itself and some kind of interaction of it with
other
covariates...

Apu Chakraborty

I have Stata 9 and wish to test a variable for putative interaction in a
linear regression model where the predictor and dependent variables are
normally distributed. Please could someone let me know how to do this.


*
*   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   |   What's new   |   Site index