[email protected]
> I have a two level data set: students nested within school. Assume a
> simple case where there are only 4 variables:
>
> Y: math achievement, the dependent variable
> X: family background meausred in SEI, individual level covariate
> Z: school - average SEI, school level covariate
> S: school ID.
>
> I want to estimate a series of simple regression of Y on X
> within each
> context (each school), extract the coefficient vectors, and
> stack them
> (coefficient vector for each school) to form a coefficient matrix to
> include coeffcient vectors for all schools. What is the
> best way to do this?
-statsby- ?
Another possibility might be -parmby-, which is part of the -parmest-
package. Type -ssc desc parmest- in Stata to find how to download the
latest Stata 8 version from SSC. Alternatively, if Shige Song does not have
Stata 8, then it is possible to download the Stata 7, 6 or 5 versions of
-parmest- from my website at
