Any possible methods to define level variables in xtreg/xteregress?

[Li Tianhao]

Dear all, Now I am working with a 3-year dataset on education. Here I want to explore how home-school communication affect students' achievements. Since the indepedent variable is exogenous, it seems that xteregress fit best here. But I should also control variables on class and school level (e.g. teachers and school resources). Are there any possible methods to define those variables?(Just like command mixed) Thanks for all help.

[Carlo Lazzaro]

Li:
if your independent variable is exogenous, you can safely use -xtreg-.
The remaining predictors are categorical (and trivially so):see -fvvarlist- notation for their creation.
Eventually, if the study design is nested (that is, pupils are nested within classrooms, which are, in turn, nested within schools, which are, in turn nested within school districts and so on, -mixed- is the way to go provided that your dependent variable is continuous.

[Li Tianhao]

Li:
if your independent variable is exogenous, you can safely use -xtreg-.
The remaining predictors are categorical (and trivially so):see -fvvarlist- notation for their creation.
Eventually, if the study design is nested (that is, pupils are nested within classrooms, which are, in turn, nested within schools, which are, in turn nested within school districts and so on, -mixed- is the way to go provided that your dependent variable is continuous.

Thank you very much, Professor. I have decided to use the nested model.
In the first I made a mistake. The independent has an endogeneity issue. Then how should I deal with the independent variable(if I still want to use the nested model with treatment effect)?

[Carlo Lazzaro]

Li:
please, call me Carlo . Thanks.
I am not aware of any official Stata command that perforrms instrumental variabe regression with nested data.
That said, you may want to take a look at Multilevel Modelling Newsletter (herts.ac.uk).

[Li Tianhao]

Li:
please, call me Carlo . Thanks.
I am not aware of any official Stata command that perforrms instrumental variabe regression with nested data.
That said, you may want to take a look at Multilevel Modelling Newsletter (herts.ac.uk).

Carlo:
Thank you very much for your detailed and immediate advice. That helps me a lot!
I will go through the article soon.

[Li Tianhao]

Li:
please, call me Carlo . Thanks.
I am not aware of any official Stata command that perforrms instrumental variabe regression with nested data.
That said, you may want to take a look at Multilevel Modelling Newsletter (herts.ac.uk).

Carlo:
I really appreciate your help very much.
After finishing that article, I also have a look on Solving endogeneity problems in multilevel estimation: an example using education production functions.
Sadly given a binary independent variable, it is still hard to cope with such an endogeneity issue. Coz logit model should be used instead of traditional linear model and it is hard to merge the estimation model and full one together.
Now I take a more conservative approach -- to display both mixed Command and xtergress Command results.