There is an article specifically addressing weighted vs. non-weighted
estimates for the NSFH. I think it is by Bumpass and Sweet, but I'm not
positive. I happen to disagree with them. I think you should always
account for the sample design. But the documentation for the sample design
of the NSFH is extremely weak.
Bryan Sayer
Statistician, SSS Inc.
[email protected]
-----Original Message-----
From: Ulrich Kohler [mailto:[email protected]]
Sent: Monday, September 23, 2002 5:06 AM
To: [email protected]
Subject: Re: st: how to compare weight and unweighted statistics
Yap Keng Loong wrote
> i like to know what commands to use if i want to
> compare weight and unweighted statistics in the
> dataset.
It depends a bit on the number and nature of the statistics you want to
compare. A general way may be to calculate weighted and unweighted
statistics
and store the results in a postfile. Use the following as a starting point
and take a look at -help post- for more informations.
--------------------------------------------------
postfile stats weight uweight using myfile, replace
foreach var of varlist myvars {
summarize `var' [weight=wvar] /* Or whatever statistic you like */
local weight = r(mean)
summarize `var'
local uweight = r(mean)
post stats (`weight') (`uweight')
}
postclose stats
use myfile, clear
gen diff = uweight - weight
list uweight weight diff
-----------------------------------------------------
regards
uli
--
[email protected]
http://www.sowi.uni-mannheim.de/lesas
*
* 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/
*
* 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/
