This is almost certainly happening because *among cases where
general==0*, there are one or more strata with only one PSU represented.
This is one reason that you want to use -svymean-'s subpop() option,
rather than the -if- qualifier.
Like this:
. svymean x y
Survey mean estimation
pweight: <none> Number of obs =
12
Strata: strat Number of strata =
6
PSU: psu Number of PSUs =
12
Population size =
12
------------------------------------------------------------------------
------
Mean | Estimate Std. Err. [95% Conf. Interval] Deff
---------+--------------------------------------------------------------
------
x | .6666667 .1178511 .3782953 .955038 .6875
y | .4604045 .0603939 .312626 .608183 .5576608
------------------------------------------------------------------------
------
. svymean y if x==0
stratum with only one PSU detected
r(460);
. gen xzero=(x==0) if x!=.
. svymean y , subpop(xzero)
Note: 3 strata omitted because they contain no subpopulation members
Survey mean estimation
pweight: <none> Number of obs =
6
Strata: strat Number of strata =
3
PSU: psu Number of PSUs =
6
Subpop.: xzero==1 Population size =
6
------------------------------------------------------------------------
------
Mean | Estimate Std. Err. [95% Conf. Interval] Deff
---------+--------------------------------------------------------------
------
y | .49303 .1126696 .134465 .851595 .991555
------------------------------------------------------------------------
------
--Nick Winter
-----------------------------------------------------------
Nicholas Winter, Ph.D. P 202.939.5343
Policy Studies Associates F 202.939.5732
1718 Connecticut Avenue, NW [email protected]
Washington, DC 20009-1148 www.policystudies.com
-----------------------------------------------------------
> -----Original Message-----
> From: Theodoropoulos, N. [mailto:[email protected]]
> Sent: Monday, October 21, 2002 2:46 PM
> To: [email protected]
> Subject: st: svy commands
>
>
> Dear statalisters,
>
> I am using the survey commands in Stata to find some means. I
> have removed all the stratums in which one psu (primary
> sampling unit) is detected. I check that there is no stratum
> with one psu for all my variables by using the svydes command.
> I don't have problems to get the invididual means for the
> individual variables "general" and "producti". However when i
> want to find the mean of the "general" variable given that
> "producti" variable is equal to zero i get the message that a
> stratum with one psu is detected. The "general" variable is
> continuous, and the "producti" variable is binary.
> Any hints will be highly appreciated.
> Thanks in advance.
>
> Here is a copy of the output that i get.
>
> svymean general
>
> Survey mean estimation
>
> pweight: empwt_nr Number of
> obs = 1727
> Strata: newstr1 Number of
> strata = 44
> PSU: newpsu1 Number of
> PSUs = 162
> Population
> size = 1200.4609
>
> --------------------------------------------------------------
> ----------------
> Mean | Estimate Std. Err. [95% Conf. Interval] Deff
> ---------+----------------------------------------------------
> ----------------
> general | 4.156863 .0044976 4.147957 4.16577 2.432631
> --------------------------------------------------------------
> ----------------
>
> . svymean producti
>
> Survey mean estimation
>
> pweight: empwt_nr Number of
> obs = 1727
> Strata: newstr1 Number of
> strata = 44
> PSU: newpsu1 Number of
> PSUs = 162
> Population
> size = 1200.4609
>
> --------------------------------------------------------------
> ----------------
> Mean | Estimate Std. Err. [95% Conf. Interval] Deff
> ---------+----------------------------------------------------
> ----------------
> producti | .3359512 .0475784 .2417331 .4301693 17.51394
> --------------------------------------------------------------
> ----------------
>
> . svymean general if producti==0
> stratum with only one PSU detected
> r(460);
> *
> * 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/
