I received a question from a collegue that completely puzzled me. He
has a regression model with an interaction effect. When he
mean-centered his variables (and recreated the interaction effects) the
test statistics (t-value and p-value) changes quite a bit. I did not
expect this to happen as mean-centering is just a linear
transformation. I recreated this behavior in the example below. Does
anyone have an idea what is causing this?
-- Maarten
*-------- begin example ----------
sysuse auto, clear
gen wXf = weight*foreign
reg mpg weight wXf
sum weight, meanonly
gen c_weight = weight - r(mean)
gen c_wXf = c_weight*foreign
reg mpg c_weight c_wXf
*--------- end example -----------
(For more on how to use examples I sent to the Statalist, see
http://home.fsw.vu.nl/m.buis/stata/exampleFAQ.html )
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting address:
Buitenveldertselaan 3 (Metropolitan), room N515
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
*
* 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/
