Date:         Fri, 14 Oct 2005 20:31:54 -0700
Reply-To:     David L Cassell <davidlcassell@MSN.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         David L Cassell <davidlcassell@MSN.COM>
Subject:      Re: % macro in proc surveyreg
In-Reply-To:  <200510132155.j9DKP1G8004697@malibu.cc.uga.edu>
Content-Type: text/plain; format=flowed

hsiehchu@MSU.EDU wrote back to me personally instead of to SAS-L:
>Hello David:
>
>thanks for your response, which does help me clear up my mind.
>however, i ran into another problem:
>in my dataset there is one strata with only one PSU present, and in this
>situation SAS cannot compute variance for me.
>do you have any suggestions on that?
>thanks in advance for your advice.
>
>
>%inc 'c:\sregsub.sas';
>%sregsub(data=startover,
>strata=strata,
>cluster=psu,
>weight=f2qwt,
>class=f2s39e2,
>model=comf2svoc=f2s40j2 f2s39e2 f2s40f2 f2s99h2,
>subpop=domain=1)
>run;

I would say that you'll have to live with this.  You can't go back out and
get more
PSU's can you?  This was just poor planning in the survey design phase.
Even tools
like SUDAAN which do jackknifing only do jackknifing between clusters at the
first
stage of sampling.

You ought to get a variance estimate.  The regression analysis method under
the hood should collapse this stratum with some other stratum that has
multiple
clusters.  In fact, it may collapse it into the preceding stratum, whether
that is a
good choice or not.  If the macro balks, then you could do this collapsing
yourself.

In the meantime, let me be even more picky.  Naming your stratum variable
STRATA
is really not informative.  It might be more helpful to others if you gave
it a better
name.

HTH,
David
--
David L. Cassell
mathematical statistician
Design Pathways
3115 NW Norwood Pl.
Corvallis OR 97330

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search!
http://search.msn.click-url.com/go/onm00200636ave/direct/01/