Re: st: Interactions in xreg

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Mon, 23/8/10, Morten Hesse <mh@crf.au.dk> wrote:
> I have data, where subjects are tested in 5 blocks.
> Subjects vary on  variable "ant" (dummy coded). I want
> to test of there is a curvi-linear effect of "ant" on
> outcome igt.
> blocksq=block^2 (to test curvi-linearity).

This really depends on what is in your variable block.
If I see a variable names block I will by default assume
that it is categorical. In that case, computing the 
square of that variable makes no sense. It could be an
ordinal variable. In that case computing the square
still makes no sense. You can only compute the square
if you believe the values that the variable block can 
take also gives you information on the distances between
the categories. Stata obviously cannot see what you 
believe, so it will alow you to compute a square of 
any variable, regardelss of whether that makes sense
or not.

If block is categorical or ordinal you must enter it
as a categorical variable, e.g.:

xtreg igt i.ant#i.block age gender, mle

assuming you have Stata 11.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   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/