Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: creating interaction variables


From   Richard Williams <Richard.A.Williams.5@ND.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: creating interaction variables
Date   Fri, 11 Jun 2004 08:55:15 -0500

At 01:38 PM 6/11/2004 +0900, Joseph Coveney wrote:
-help xi- will show how to do you what you want.  You can probably automate
the exercise, too, with a couple of -foreach- or -for- loops.

You'll need to stay aware whether one or both the predictor variables are
continuous in each instance of the many different interaction terms that
you're exploring, since -xi-'s syntax will be slightly different from the
case in which they are both categorical.
Doesn't -xi- require that one of the vars be categorical though? How about if both x's are continuous? If it doesn't exist already, it seems like it would be nice to have something like -xi- for continuous vars, e.g. if you have x1*x2 it expands to include the terms x1, x2 and x1*x2, so you don't have to bother computing the interactions beforehand.

As an alternative, I think -anova- might do the trick, e.g. you can have syntax like

anova y x1 x2 x1*x2, continuous(x1 x2)

As a possibly unnecessary sidelight, going back to the original requested regression:

regress y x1*x2 x3 x4

However the problem gets solved, you'd probably want to have the lower order terms included, e.g. it should expand to

regress y x1 x2 x1*x2 x3 x4


*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* 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