Re: st: about multi-stage stratified sampling designing

[Steve Samuels <sjsamuels@gmail.com>]

I second Stas's request for the SUDAAN design statements. In SUDAAN,
unlike Stata, design statements are included in a block with the
statistical procedure statements.

Be sure to tell us what version of Stata you are using.  Also, if
there is a web reference in English to the survey design, please give
it. If you did not know how to -svyset- your data, then I am not
surprised that your Stata and SUDAAN results differed.

Note that our program's name is "Stata", not "STATA".   In English,
names are upper case only if the letters stand for other words in the
original name.  For example "SAS" = "Statistical Analysis System".
SUDAAN = "Survey Data Analysis".  SPSS = "Statistical Package for the
Social Sciences".

Steve

On Mon, Apr 26, 2010 at 9:30 AM, Stas Kolenikov <skolenik@gmail.com> wrote:
> On Mon, Apr 26, 2010 at 4:51 AM, ęę­ē´ <sherinalee@gmail.com> wrote:
>> Because ofĀ research, I am working a database which is multi-stage
>> stratified systematic sampling designing. Specially, it adopted three
>> kinds of multi-stage stratified sampling designing that I couldnā€™t use
>> the normal STATA command- svyset. And how can I do it?
>>
>> By the way, The database managers suggested that users handle the
>> weight by SUDAAN(statistical software) and they released the examples
>> of result. There was a post-stratification weight so that I tried to
>> use the STATA command-poststrata. Unfortunately the results differed
>> from using SUDAAN. SUDAAN declared the sampling designing(strata, psu)
>> and post-weight and the way of STATA used the post-stratification and
>> post-weight.Ā  Could you tell me what the different between them?
>
> What are the SUDAAN design specification that were suggested? I don't
> think there's anything left in SUDAAN that Stata cannot do.
>
> With plain vanilla systematic sampling, there is no way to estimate
> the variance. That is to say, you can use any design statements in
> Stata, but none of them will give meaningful estimates. One needs to
> use some special tricks at the design/sampling stage, such as using
> several seed/threads of sampling to create an effective 2
> interpenetrating PSU/stratum design.
>
> --
> Stas Kolenikov, also found at http://stas.kolenikov.name
> Small print: I use this email account for mailing lists only.
>
> *
> *   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/
>



-- 
Steven Samuels
sjsamuels@gmail.com
18 Cantine's Island
Saugerties NY 12477
USA
Voice: 845-246-0774
Fax:    206-202-4783

*
*   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/