----- Original Message -----
From: "Vimalanand S. Prabhu" <[email protected]>
To: <[email protected]>
Sent: Wednesday, April 16, 2003 11:01 PM
Subject: st: SUR on stacked data
> Dear Friends,
>
> I have data stacked into columns. I want to conduct a SUR on this data
where
>
> Y1 = f(x1)
> Y2= f(x2)
> where y1, x1 and y2, x2 are variables for different gender.
> Note that both Xs are not same (else SUR is algebriacally equivalent to
> OLS).
>
> Do you know how to convert a stacked dataset into a formal where I can
> directly use "sureg" command for SUR model?
>
> Thanks,
>
> Prabhu
>
Prabhu,
I believe the following little program will work. It creates a series of
new variables for each group beginning at the same observation number as the
first group.
Example:
use http://www.stata-press.com/data/r8/grunfeld.dta
separate2 invest mvalue kstock, group(company)
and then
sureg (invest1 mvalue1 kstock1) (invest2 mvalue2 kstock2) ....
I hope this helps,
Scott
-----start of separate2.ado-----
program separate2
version 8.0
syntax varlist, group(varname)
tempname Vals /*Pilfered from Nick Cox's -levels- */
qui tab `group' , matrow(`Vals')
local nvals = r(r) -1
local sep " "
forval i = 1 / `nvals' {
local val = `Vals'[`i',1]
if `i' < `nvals' local vals "`vals'`val'`sep'"
else local vals "`vals'`val'"
}
local levels `"`vals'"'
qui sum `group'
local min = r(min)
foreach var of varlist `varlist' {
local i = `min'
local num = 0
separate `var', by(`group')
foreach x of local levels {
qui sum `var'`i'
local num = `num' + r(N)
local i = `i' + 1
replace `var'`i' = `var'`i'[_n+`num']
}
}
end
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
