Re: [R] bootstrapping quantile regression

2012-10-31 Thread David Freedman
A possiblie solution might be to use the survey package. You could specify that the data is clustered using the svydesign function, and then speciy the replicate weights using the as.svrepdesign function. And then, it would be possible to use the withReplicates function to bootstrap the clusters

Re: [R] bootstrapping quantile regression

2012-10-31 Thread Frank Harrell
A piece of this is solved by the rms package's Rq and bootcov functions. -Frank Roger Koenker-3 wrote > There is no automatic "clustering" option for QR bootstrapping. > You will have to roll your own. > > > url:www.econ.uiuc.edu/~rogerRoger Koenker > email > rkoenker@ >

Re: [R] bootstrapping quantile regression

2012-10-31 Thread Roger Koenker
There is no automatic "clustering" option for QR bootstrapping. You will have to roll your own. url:www.econ.uiuc.edu/~rogerRoger Koenker emailrkoen...@uiuc.eduDepartment of Economics vox: 217-333-4558University of Illinois fax: 217-244-66

Re: [R] bootstrapping quantile regression

2012-10-30 Thread Kay Cichini
sry, I forgot to replace rlm() - but actually I tried both and the question applies to both approaches.. Am 31.10.2012 00:19 schrieb "Kay Cichini" : > > HI everyone, > > I try to get some bootstrap CIs for coefficients obtained by quantile regression. I have influencial values and thus switched to