This wisdom is shared by others. See comments embedded
within
McCullagh, P. and Nelder, J.A. 1989. Generalized linear
models. London: Chapman and Hall.
on whether the origin is special.
I add a small comment on the original formulation.
This is essentially a statistical question, and not
an econometric question.
Nick
[email protected]
Mirko
> I really invite you to have a look at this web page by Matt Golder
> full of hints about multiplicative interaction terms, correct standard
> errors and Stata codes, which is a companion of:
>
> Thomas Brambor, William Roberts Clark, Matt Golder
> "Understanding Interaction Models: Improving Empirical Analyses"
> (Political Analysis 14: 63-82),
>
> http://homepages.nyu.edu/~mrg217/interaction.html
>
> Regarding your point and in brief, they say:
>
> "We recommend that scholars include all constitutive terms in their
> interaction model specifications".
Richard J Stoll
> > On Wed, 23 May 2007, Wanli Zhao wrote:
> >
> > > I have a basic(?) econometric question on regressions
> with interactions. My
> > > search on this did not product much. Do I have to include
> the linear terms
> > > in the regression when I also have interactions (include
> x & y when I have
> > > x*y)? What does it mean when I drop x, or y or BOTH
> (always keep x*y)? Any
> > > help is highly appreciated (occasions where different
> model is suitable),
> > > especially what I should get to read!
> > >
> >
> > Yes.
> >
> > The "classic" article in political science is:
> > Friedrich, Robert. 1982. "In Defense of Multiplicative
> Terms in Multiple
> > Regression Equations." American Journal of Political
> Science 26: 797-833.
*
* 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/
