Re: st: missing standard errors with single sampling units-help

[Steven Samuels <sjhsamuels@earthlink.net>]

Adbola, you misunderstand the error message.  You must combine  
adjacent strata, not adjacent sampling units.  As Joao stated, you  
must -svyset- your data.  You have not shown us the -svyset-  
command.  And, look at -svydes-.  You have, in your second listing,  
asked for final stage units, but the error is really an error for  
first stage units, the default shown by -svydes-, if you do not  give  
the "finalstage" option.

See the email message that Joao referred to for another suggestion  
besides combining strata. If you do not show us exactly what you  
typed  (starting from -svyset- and -svydes- )and exactly what Stata  
responded, we cannot give you much help. (see the Statalist FAQ,  
especially section 3). If you are unsure of the survey concepts, then  
consult Sharon Lohr's book  "Sampling: Design and Analysis", Duxbury  
Books. At Harvard, Professor Alan Zaslavsky in the Medical School is  
an excellent resource.

-Steve




On Sep 10, 2008, at 1:35 PM, Joao Ricardo F. Lima wrote:


> Dear Adebola,
>
> see this Steven�s message:
>
> http://www.stata.com/statalist/archive/2008-08/msg01377.html
>
> Do you svyset your data again with the new stratum variable?
>
> I think that you can use idonepsu to do this too...
>
> help idonepsu
>
> Hope this helps,
>
> Joao Lima
>
> 2008/9/10 Adebola Odunlami <aodunlam@hsph.harvard.edu>:
> > Hi all,
> >
> > I am seeking help to a problem with trying to get appropriate  
> > numbers for my standard errors.
> >
> > So here's the problem. When I used weights to assess means of a  
> > variable, mpsyall6nodrm
> > by race( I have 3 categories), I don't get any standard errors for  
> > my means. Instead I get a message, stating "missing
> > standard errors for weighted sample because of stratum with single  
> > sampling unit." Someone told me to try to see which stratum has a  
> > single sampling unit in hopes of combining them. I found 5 single  
> > sampling units, combined them with ones preceding them but this  
> > still doesn't fix the problem when I check it again. What am I  
> > doing wrong and how can I fix this?
> >
> > Commands:
> > svydes mpsyall6nodrm
> > gen jhumka = clust
> > recode jhumka (29 = 30) (31= 32) (46 = 47) (88 = 89) (91 = 90)
> > svydes mpsyall6nodrm, finalstage
> >
> >
> > Adebola Odunlami, MPH
> > Doctoral Student
> > Society, Human Development and Health
> > Harvard School of Public Health
> > 301 256 4655
> > *
> > *   For searches and help try:
> > *   http://www.stata.com/help.cgi?search
> > *   http://www.stata.com/support/statalist/faq
> > *   http://www.ats.ucla.edu/stat/stata/
>
>
> --
> -------------------------------
> Joao Ricardo Lima
> Professor
> UFPB-CCA-DCFS
> +553138923914
> -------------------------------
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/