[R] Complex sample variances

[Thomas Lumley]

On Wed, 14 Apr 2004, Baskin, Robert wrote:

>
> This method is basically randomly building BRR replicates (in a 2-per
> design) so it is like an inefficient BRR and the number of bootstrap
> replicates needed may depend on both the statistic being estimated and the
> number of replicates in a fully balanced BRR set.
>

And the survey package does already have JK1, JKn, BRR, and Fay's method:
you can convert a survey.design object with design information to a
svrep.design object with replicate weights and proceed a la WesVar.

I've added a svrepquantile function, which will be in a new release soon
-- I need to finish revising it for JSS.

	-thomas