Re: st: difference between interaction created with # and created with xi3

["Michael N. Mitchell" <Michael.Norman.Mitchell@gmail.com>]

Dear Uso

  The short answer to your question is that this xi3: command...

xi3: regress y i.a*i.b*i.c

  is the equivalent of

regress y i.a##i.b##i.c

  which is the equivalent of

regress y i.a i.b i.c ///
          i.a#i.b i.a#i.c i.b#i.c ///
          i.a#i.b#i.c

  On the other hand, specifying

regress y i.a#i.b#i.c

  yields a model that only specifies the interactions (no main effects, no two way 
interactions), which is not what we generally desire. I hope it suffices to say that this 
is not a particularly interesting model, and is complicated to explain via email.

I hope that helps,

Best regards,

Michael N. Mitchell
Data Management Using Stata      - http://www.stata.com/bookstore/dmus.html
A Visual Guide to Stata Graphics - http://www.stata.com/bookstore/vgsg.html
Stata tidbit of the week         - http://www.MichaelNormanMitchell.com



On 2010-08-23 12.49 PM, uso stata wrote:
> Hello statlist,
> I am currently using a linear model specification and I am fiddling
> with interactions.
>
> I have three BINARY regressors a b c
>
> When I use the command a#b#c stata spits out 7 interactions:
>
> a b c
> 0 0 1
> 0 1 0
> 0 1 1
> 1 0 0
> 1 0 1
> 1 1 0
> 1 1 1
>
> when I use xi3: and a*b*c I get:
> a
> b
> c
> _Iaxb
> _Iaxc
> _Ibxc
> _axbxc
>
> While the main effects are identical, i.e. the coefficient of "0 0 1"
> is the same as the coefficient of "c", the interaction terms are all
> different.
>
> P
>
> The xi3 methods generates the same coefficient as the a##b##c method,
> that is equivalent to i.a i.b i.c a#b#c
>
> So my question is what is the specification that stata creates for
> a#b#c? How do I interpret for example "0 1 1", I thought it was the
> same as the interaction of b and c, but it seems not.
> *
> *   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/
*
*   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/