Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: decomposing nested fixed effects

From   "Lim, Raymond" <>
To   "" <>
Subject   st: decomposing nested fixed effects
Date   Tue, 8 Nov 2011 15:22:17 -0500

Hello Statalisters,
Is there a method for decompose fixed effects of a nested variable into effect due to the lower level vs. effect due to the higher level? For example, say we are studying student test scores across time, in different schools. Students are nested in schools (i.e. students do not appear in different schools). I run a regression with student fixed effects; is there a way to decompose this fixed effect and say 40% is due to the student and 60% is due to being in a particular school?

I can't run with school dummies because I already have student fixed effects that perfectly identify the school. My research leads me to multilevel modeling and xtmixed, but it seems to require treating the effect as random. However, I think the unobserved effects are correlated with the other covariates, thus requiring a fixed effect regression. For example, school effects is correlated with classroom spending, student effects is correlated with time spent studying, etc.

Raymond Lim
Research Coordinator
Columbia Business School - Finance & Economics Division
*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index