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

st: RE: Using more than one weights variable in svylogit?


From   Lee Sieswerda <Lee.Sieswerda@tbdhu.com>
To   "'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Using more than one weights variable in svylogit?
Date   Mon, 16 Jun 2003 19:24:52 -0400

Your choice depends on how well you know your sampling frame. It sounds to
me like you probably know the composition of the agency sampling frame, and
so can correct for over/under-representation of each class of agency. Do you
also know the sex composition of your sampling frame? If so, then there
should be no problem making a composite weighting variable.  
e.g.,
Sample proportions
	sex: Male = .1
	agency: Dist = .8
Actual proportions
	sex: Male = .3
	agency: Dist = .5

post-strat weights:
	Male-Dist: (.3/.1)*(.5/.8)
	Male-Reg: (.3/.1)*(.5/.2) 
	Fem-Dist: (.7/.9)*(.5/.8)
	Fem-Reg: (.7/.9)*(.5/.2)

If you do not know the sex distribution of the sampling frame, then you
probably should not try to weight on sex.

Regards,
Lee

Lee Sieswerda, Epidemiologist
Thunder Bay District Health Unit
Lee.Sieswerda@tbdhu.com


-----Original Message-----
From: Julia Koschinsky [mailto:pin.ny@albany.edu] 
Sent: Monday, June 16, 2003 3:05 PM
To: 'statalist@hsphsun2.harvard.edu'
Subject: st: Using more than one weights variable in svylogit?


Dear Stata List,

We are trying to correct for the overrepresentation of women in our sample.
We also have an overpresentation of redistributive and distributive agencies
(as opposed to regulatory agencies) in the sample. We created two inverse
weights variables, one for gender and one for agency type, that we would
like to use in a logit model.

We have run the svyset pweights and svylogit commands with the gender
weights variable and are wondering:

1.Is it possible to run both the gender and agency weights variable in the
svylogit command? 2. Or would it make sense to create a combined
gender-agency weights variable (if so, should we sum or multiply the
weights)? 
3. Or is it sufficient to only include the gender weights variable since
many of the female respondents also work in redistributive agencies?

Many thanks,
Julia Koschinsky
University at Albany
*
*   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/



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