elhadjilaouansidi@vandals.uidaho.edu wrote to me personally instead of to SAS-L: > >Sir, >I would like to generate 500 bootstrap weights based on a complex design, >say, a stratified random sampling. >I would appreciate any suggestion. >Thanks >Elhadji Sidi >MS Statistics

First, always write to SAS-L isntead of to me personally. You ended up in my spam filters, and it's a miracle I even saw this today.

Second, don't be doing boostraps on complex sampel designs unless you're fully conversant in the technical details of design-adjusted bootstraps. It's not trivial. You cannot use standard bootstrap estimation when the data are not exchangeable, as in survey sampling.

Third, standard survey sampling estimation techniques are doing things like Taylor series linearization or jackknifing, os bootstrapping that is like bootstrapping your bootstrap. It's not worth the effort.

Perhaps if you write back to SAS-L - and not to me personally unless you want to hire me as a consultant - and explain your problem and your data and your reason for wanting a bootstrap at all, then someoen here might be able to help you out.

HTH, David -- David L. Cassell mathematical statistician Design Pathways 3115 NW Norwood Pl. Corvallis OR 97330

_________________________________________________________________ Hotmail to go? Get your Hotmail, news, sports and much more! http://mobile.msn.com