Michael, You have not described your design with enough detail for us
to know how to -svyset- your data. So that we can assist you, please
answer the following questions.
• What is the purpose of the study? What analyses are you planning?
Are you testing hypotheses? ? Do you plan to run regression models?
• What will be the units of analysis: people? households? both?
• Was an independent sample of ea's drawn from each province? OR, was
there a single sample of n = 340 ea's drawn at one time from the
pooled set of ea's in the 20 provinces?
• Was one sample taken from urban ea's and another from rural ea's?
OR. Was the original sample of ea's drawn from the population of all
ea's without regard urban/rural status. (If you sampled from provinces
independently, add "within each province" to the end of these
questions.)
• Was the sampling with replacement at any stage?
• "do I need another variable -n_hh2- equal to the number of hhs per
urban/rural location?" What do you mean by "location" here? Do you
have urban/rural information for households different from the
urban/rural classification of their enumeration area?
• Were there any non-responding households?
-Steve
On Wed, Jan 13, 2010 at 2:54 AM, Michael Begg <[email protected]> wrote:
> Hi,
>
> I am trying to set the survey design for the following survey:
>
> A two stage stratified sampling procedure based on provinces and urban/rural location
>
> 20 provinces were chosen for the survey, from which, 340 enumeration areas were randomly selected, from which, 30 households were randomly sampled if the EA was urban and 40 households were randomly sampled if the EA was rural.
>
> I am unsure what the syntax for the svyset command should be in the case of the 2 stratum
>
> If I generate the following variables:
> province = a indicator of the provinces chosen for the survey;
> ea = an indicator of the EA chosen for the survey;
> rural = an indicator of urban/rural location.
> n_ea = the number of EAs per province.
> n_hh = the number of hhs per ea
>
> I am on the right track with what's below or do I need another variable -n_hh2- equal to the number of hhs per urban/rural location?
>
> svyset ea [pw=weight], strata(province) fpc(n_ea) || hh, strata(rural) fpc(n_hh)
>
> Cheers
>
> Michael]
--
Steven Samuels
[email protected]
18 Cantine's Island
Saugerties NY 12477
USA
845-246-0774
*
* 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/
