Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: The use of pweights with regress


From   "Clive Nicholas" <Clive.Nicholas@newcastle.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: The use of pweights with regress
Date   Fri, 5 Nov 2004 19:27:24 -0000 (GMT)

Barbara Hamilton wrote:

[...]

> would my standard error estimates be compromised as Winship and Radbill
> noted back then, or does pweight now correct for any problems with these
> estimates brought about by the act of weighting?

Without pretending that I can demonstrate this algebraically (over and
above what you've already found in the Stata manuals and on-line help),
the use of -pweight-s act to correct for any sample selection biases that
may arise from your data.

> Also, can anyone point me to any documentation as to exactly what/how
> pweights are applies.  For example, I know that fweights are analogous
> to running an unweighted regression on a dataset that has the correct
> number of observation replications denoted by the fweight.  However, I
> don't have a concept in my head of how pweights are actually
> applied...that is, can anyone briefly state a conceptualization such as
> the one I made above about pweights, or does anyone have access to
> technicalities/formulas about how pweights are applied?

Normally, probability weights weight all observations in a dataset by
producing -pweight-s equal to 1/(probability of selection) (although you
can change this to whatever you think is appropriate). So, if a black,
middle-class male living in the South of England has only a probability of
being selected for a UK political opinion survey of 0.05, his -pweight- is
(1/0.05) = 20.

The Winship and Radbill (1994) paper that you refer to strongly recommends
the use of the DuMouchel-Duncan (1983) F test to test the significance of
the impact of sampling weights on estimation results. Ben Jann has
implemented this post-estimation test for Stata with the -wgttest-
package. You can download this from SSC:

. ssc desc wgttest
. ssc inst wgttest

Hope all that helps.

CLIVE NICHOLAS        |t: 0(044)7903 397793
Politics              |e: clive.nicholas@ncl.ac.uk
Newcastle University  |http://www.ncl.ac.uk/geps

DuMouchel WH and Duncan GJ (1983) "Using Sample Survey Weights in Multiple
Regression Analysis" J AMER STAT ASSOC 78(383): 535-43.
Winship C and Radbill L (1994) "Sampling Weights and Regression Analysis"
SOCIOLOGICAL METHODS AND RESEARCH 23(2): 230-57
*
*   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