Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Maarten buis <maartenbuis@yahoo.co.uk> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Interactions in xreg |
Date | Mon, 23 Aug 2010 10:33:11 +0000 (GMT) |
--- On Mon, 23/8/10, Morten Hesse wrote: > I did that too, and it works. "Block" is of course an > ordinal variable (1-5). Yet, I want to test whether > the group with characteristic "ant" has a different > pattern of behaviour over "blocks", and I have reasons > to believe that that pattern is curvilinear, compared > with subjects who do net have characteristic "ant" > (scores increases towards the middle, of the test, > decreases towards the end). I can of course test > whether the parameter for subjects during the > different blocks are significant (e.g., start versus > middle, and middle versus start). And indeed, they > are. But I have never seen that used as a test of > curvilinearity, and I feel that it is a bit "informal". > Am I wrong? Yes, unfortunately you are wrong. Such quadratic terms have no meaning in an ordinal variable. If you do so you are basically assuming that the distances between your 5 categories are equal. -- 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/