Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: tabulate with pweight


From   "Austin Nichols" <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: tabulate with pweight
Date   Mon, 19 Mar 2007 11:35:19 -0500

Eva--
I will assume you do in fact have a pweight on your data. While it is
quite true that "use of pweights generally requires prior use of
-svyset- and then -svy- commands" as Nick points out, note that
specifying pweights as aweights produces identical point estimates, so
*as long as you are not computing the variance* of estimates, for
example to test whether proportions are equal or to compute confidence
intervals, just:
tab x y [aw=weight]
and you can see the two weight types side by side by trying:
reg x y [aw=weight]
reg x y [pw=weight]
where only the Std.Errors and related quantities should differ.

On 3/19/07, n j cox <n.j.cox@durham.ac.uk> wrote:
It is not clear to me that your situation is one leading
to the use of pweights.

Either way, -tabulate- does not support pweights. That
is implied in the help, which explains that only the
other kinds of weight are supported.

Use of pweights generally requires prior use of -svyset-
and then -svy- commands.

Nick
n.j.cox@durham.ac.uk

Eva Gottschalk <e-gott.81@web.de>

I'm using data that is weighted for the overpresentation of east-germany
(weighting variable=weight).

For the regression I just add [pweight=dweight] to weight my data.

This command does not work with: tab x y [pweight=weight]

I don't understand why, because it seams quite important to wheight the
data for descriptions about germany as a whole.
*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index