--- [email protected] wrote:
> I'm interested in creating a mixed model with two predictors (x and
> y) and dependent variable (z). My data has couples (males and
> females).
> Would the mixed model look like this:
> xtmixed z x y || couple: gender
> I'm not concerned with the covariance structure at the moment.
> x, y, z are continuous values. I want to do a regression of x,y on z
> where
> couple gender
> 1 1
> 1 2
> 2 1
> 2 2
> 3 1
> 3 2
That still depends on what you want to model, at the very least you
probably either want:
xtmixed z x y gender || couple:
or:
xtmixed z x y gender || couple: gender
In the syntax you gave (which is strictly speaking legal) you forced
the distribution of the parameter of gender to be centered around zero
(i.e. you forced gender to have on average no effect, but in individual
households that could differ). I would start with the first model (the
one with gender in the fixed effects but not as a random variable).
If you want to some more reading on this subject I can strongly
recommend "Multilevel and Longitudinal Modeling Using Stata", 2nd
Edition by Sophia Rabe-Hesketh and Anders Skrondal that appeared in
2008 with the Stata-Press: http://www.stata-press.com/books/mlmus2.html
Hope this helps,
Maarten
-----------------------------------------
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/
