David Binder and Georgia Roberts of Statistics Canada gave an
excellent talk the last JSM on approaches to complex survey data
analysis, with the main focus on (mis)match between the survey
population and the analysis population. See if you can find the ASA
proceedings CD-ROM with their short paper, and the real paper should
be coming out soon, too, I believe.
Technically, you would want to designate each of your population as strata:
egen superstrata = group(population strata)
svyset , strata(superstrata)
and then proceed as usual.
On 5/21/07, Hillel Alpert <HALPERT@hsph.harvard.edu> wrote:
Could someone advise regarding analysis with Stata of multi-population data with complex sample designs?
The data are from a standardized survey administered separately in multiple populations. For each population, the data are weighted to adjust for sample selection ((school and class levels), non-response (school, class, and student levels), and post-stratification of the sample population relative to the grade and sex distribution in the total population. Thus, in each population data set, weight, strata, and psu variables are associated with each observation.
Could someone help with the question, how should the weighting and stratification be done in Stata to analyze the data across populations, i.e for cross-population comparisons and for aggregate computations). What additional information is needed regarding the populations (population sizes)?
Thank you for recommendations.
Hillel Alpert
halpert@hsph.harvard.edu
*
* 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/
--
Stas Kolenikov
http://stas.kolenikov.name
*
* 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/