Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Austin Nichols <austinnichols@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Quantile regression and weights |
Date | Fri, 25 Jun 2010 15:48:53 -0400 |
Jacob Felson <felsonj@gmail.com> : Don't pretend you have freq weights when you don't. You can manually bootstrap by repeatedly resampling and estimating the qreg with aweights, which works well in practice but takes a long time to run. See http://www.stata.com/statalist/archive/2007-09/msg00147.html http://www.stata.com/statalist/archive/2009-04/msg00271.html http://www.stata.com/statalist/archive/2009-11/msg00901.html etc. on qreg, and e.g. http://www.stata.com/statalist/archive/2006-12/msg00471.html on the "manual bootstrap" idea. On Fri, Jun 25, 2010 at 3:05 PM, Jacob Felson <felsonj@gmail.com> wrote: > Hello, > > I am wondering whether there is any way to run quantile regression > (qreg) with probability weights. I understand that this cannot be > done directly, but I wondered if a work-around solution was advisable. > (Namely, weighting the data beforehand.) I tried to weight using the > analytical weight feature, bur ran into convergence problems when > trying to estimate a 0.1 quantile regression. My dataset is a subset > of about 600K cases from the 2008 Public Use Microsample (PUMS) from > the US Census. The weight I am trying to use is a probability weight > and has a very large range. What about using frequency weights with > bootstrapped standard errors? * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/