On Mon, Nov 26, 2012 at 5:43 PM, JVerkuilen (Gmail) wrote:
> I'd even argue that centering categorical predictors is often useful,
> though that's not done as often. This is especially true for variables
> not primarily of substantive interest.
For that I typically prefer what is sometimes called "effect coding",
which means all categories are compared to the grand mean. One example
of a situation where this often makes sense is when your categories
are provinces or states; that way you compare those provinces with the
national average rather than with some arbitrary reference province.
In Stata 12 you can use the -contrasts- command with the -g.- or -gw.-
operators.
-- Maarten
---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany
http://www.maartenbuis.nl
---------------------------------
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
