----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Wednesday, January 29, 2003 2:56 AM
Subject: st: -xtdata- + retain untransformed variables?
> Thank you Scott,
>
> I found where my problem was. Another little question concerning xtdata,fe:
>
> When I use it on the panel data set I want to regress afterwards, the <year>
> variable is transformed, too. I can recover the old values using -int-, but
> this does not seem the safest way to me. Is there a way of applying -xtdata-
> to only a part of the dataset, and to retain some other variables (such as
> <year>) in it?
No. However you could transform that data, save it as another data set and then
merge to obtain the year variable.
Or you could do the transformation by hand, such as:
clear
use "C:\Stata8\grunfeld.dta", clear
*Create group means
bysort comp:egen inv_m = mean(inv)
bysort comp:egen mva_m = mean(mva)
*Create overall mean
egen mva_m2 = mean(mval)
egen inv_m2 = mean(invest)
*Demean variables
gen de_inv = invest - inv_m + inv_m2
gen de_mva = mvalue - mva_m + mva_m2
reg de_inv de_mva
*Compare to xtreg, fe
xtreg invest mval,fe
Keep in mind, that your standard errors will be too small. Is there a reason
you cannot use dummy variables?
Scott
*
* 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/
