I have a sample design that can be resumed in:

  1. Split the city in X areas and, for each area, sample 5% of the residences and get some informations about it;
  2. For each residence, interview ALL residents;
  3. In the end, weights to homes and residents are calculated.

So, actually, I have 2 samples: one for residences and another for residents. My question is: what are the sampling designs and how declare it on survey()?

Each dataset has this variables: CD_APONDE - A code to identify the splited areas V0300 - A code to identify the residences V0010 - The weight

The first one (for the residences), I believe that is a stratified sampling and should declare like svydesign(strata = ~ CD_APONDE, weights = ~ V0010, data = dados).

The second (for residents, and my true question) I guess that is a stratified 1 - level Cluster Sampling design, but have not sure, and should use svydesign(ids = ~ V0300, strata = ~ CD_APONDE, weights = ~ V0010, data = dados).

Am I right? Wrong? 50/50?

  • $\begingroup$ ats.ucla.edu/stat/mult_pkg/faq/svy_howtochoose.htm might help $\endgroup$ – Anthony Damico Apr 14 '14 at 15:10
  • $\begingroup$ Thanks @AnthonyDamico, but my problem isn't only in survey(), but is also in identify the sample design correctly. I know that in all cases the pontual estimation will be the same (at least, for the many I tried), but I'm also interested in the standard error estimation. $\endgroup$ – Rcoster Apr 14 '14 at 16:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.