Resampling variance estimation for complex survey data
Editors’ note. This article was submitted and accepted before the new svy bootstrap prefix was made available in the Stata 11.1 update. The variance estimation method implemented in the new svy bootstrap prefix is equivalent to the one in bs4rw. The only real difference is syntax. For example,
. bs4rw, rw(bw*): logistic highbp height weight age female [pw=finalwgt]is equivalent to
. svyset [pw=finalwgt], vce(bootstrap) bsrweight(bw*) . svy: logistic highbp height weight age femaleSimilarly, the example using mean bootstrap replicate weights,
. local mean2fay = 1-sqrt(1/10) . svyset [pw=finalwgt], vce(brr) brrweight(bw*) fay(`mean2fay´) . svy: logistic highbp height weight age femaleis equivalent to
. svyset [pw=finalwgt], vce(bootstrap) bsrweight(bw*) bsn(10) . svy: logistic highbp height weight age femaleThe weights created by the bsweights command discussed in this article are equally applicable with the bs4rw command and with the new vce(bootstrap) and bsrweight() options of svy and svyset.
View all articles by this author: Stanislav Kolenikov
View all articles with these keywords: bsweights, balanced repeated replication, balanced bootstrap, bootstrap, complex survey data, Hadamard matrix, half-samples, jackknife, resampling, weighted bootstrap, mean bootstrapDownload citation: BibTeX RIS
Download citation and abstract: BibTeX RIS
Contact email@example.com if you have questions about the Stata Journal.