if restrictions will generally produce incorrect variances, though point
estimates will still be correct. However, missing values are not (at least
I hope not) treated with if restrictions. The only information needed on
unused observations to correctly calculate variances is PSU and strata.
Stata, however, is further restrictive and requires that subpop() and
missing not generate strata with single PSUs. This is where manipulation of
subpop() and missing can be handy, though I can't say that I have tested
every possibility. This best bet is to work through some trial examples by
hand to see if you can get the correct results.
Bryan Sayer
Statistican, SSS Inc.
-----Original Message-----
From: L Allen
To: [email protected]
Sent: 7/23/03 11:21 AM
Subject: RE: st: RE: RE: Svytab subgroup analysis with more than two
subgroups
Doesn't this have the same effect as using an if restriction? If so,
the
manual warns that Stata will not produce correct statistics and variance
estimates?
Richard Clerkin
Ph.D. Student
Indiana Nonprofit Project
Indiana University-Bloomington
>From: "Sayer, Bryan" <[email protected]>
>Reply-To: [email protected]
>To: "'L Allen '" <[email protected]>,
>"'[email protected] '" <[email protected]>
>Subject: RE: st: RE: RE: Svytab subgroup analysis with more than two
>subgroups
>Date: Wed, 23 Jul 2003 09:34:56 -0400
>
>This would appear to be one of the negative side effects to how Stata
>handles subpop(). But you might try setting the value of the group var
to
>missing for values outside of the subpop(). In fact, manipulating
missing
>values is a general fix for many problems with subpop().
>
>Bryan Sayer
>Statistician, SSS Inc.
>
>-----Original Message-----
>From: L Allen
>To: [email protected]
>Sent: 7/23/03 9:14 AM
>Subject: Re: st: RE: RE: Svytab subgroup analysis with more than two
>subgroups
>
>That is what I want. However, the problem I ran into when constructing
>the
>subgroup, both previously and as suggested below, is that the 3 groups
>not
>in the subgroup are still levels in the groupvar and generate a 0 count
>in
>half the cells in the table. With zeros in the marginals, a chi-square
>cannot be computed.
>
>Richard Clerkin
>Ph.D. Student
>Indiana Nonprofit Project
>Indiana University-Bloomington
>
>
> >From: Lee Sieswerda <[email protected]>
> >Reply-To: [email protected]
> >To: "'[email protected]'"
<[email protected]>
> >Subject: st: RE: RE: Svytab subgroup analysis with more than two
>subgroups
> >Date: Mon, 21 Jul 2003 09:31:24 -0400
> >
> >I'm not sure that I completely understand your problem either, but
> >something
> >like this:
> >
> >tab groupvar, gen(group)
> >gen dichot = (group1==1 | group2==1 | group3==1)
> >svytab groupvar catvar, subpop(dichot)
> >
> >would give you a chi-square test for the association between groupvar
>and
> >catvar for three out of your six groups. Is that not what you are
>looking
> >for?
> >
> >Lee Sieswerda, Epidemiologist
> >Thunder Bay District Health Unit
> >[email protected]
> >
> >
> >-----Original Message-----
> >From: L Allen [mailto:[email protected]]
> >Sent: Monday, July 14, 2003 1:56 PM
> >To: [email protected]
> >Subject: st: Svytab subgroup analysis with more than two subgroups
> >
> >
> >Dear Statalist,
> >
> >We have survey data with 39 strata. We have categorized the
>respondents
> >into 6 groups. This grouping was not used in the sampling design (it
is
>not
> >one of our stratifying variables). We want to do a svytab on 3 of
the
>6
> >groups against responses to a categorical variable; however, the
survey
> >manual warns us not to us an if restriction to select these three
>groups
> >because it will not produce correct statistics and variance
estimates.
> >However, the subpop() option does not appear to meet our needs
either
> >because we want to test the association among these three groups.
Any
> >suggestions on how to proceed would be appreciated.
> >
> >Richard Clerkin
> >Ph.D. Student
> >Indiana Nonprofit Project
> >Indiana University-Bloomington
> >
> >_________________________________________________________________
> >The new MSN 8: smart spam protection and 2 months FREE*
> >http://join.msn.com/?page=features/junkmail
> >
> >*
> >* 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/
> >*
> >* 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/
>
>_________________________________________________________________
>MSN 8 helps eliminate e-mail viruses. Get 2 months FREE*.
>http://join.msn.com/?page=features/virus
>
>*
>* 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/
_________________________________________________________________
Protect your PC - get McAfee.com VirusScan Online
http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963
*
* 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/
