Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Steve Samuels <sjsamuels@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: BRR for Quantile Regression using bs4rw |
Date | Tue, 24 Sep 2013 20:33:28 -0400 |
" -pweight- and -iweight- support were added to -qreg- in Stata 13 " Good news, Jeff. I should really reread -whatsnew- . Steve On Sep 24, 2013, at 6:08 PM, Jeff Pitblado, StataCorp LP wrote: Mustafa Hussien <mhussei4@uthsc.edu> is using -bs4rw- with -qreg- and is getting an error message about weights not allowed: > I have been working on a quantile regression model for complex survey data > (MEPS dataset) and was trying to use the user written command "bs4rw" to > do a balanced repeated replication estimation of the SEs out of qreg. Here > is my code: > > bs4rw, rw(brrrwt*): qreg $depvar $demo if subpop==1 [pw=perwt5], q(.75) > > This should work as an alternative to svy brr estimation, which is not > allowed with qreg, and in theory this is better than doing a regular > bootstrap. Although the help file for bs4rw says pweights are allowed in > command, I always get an error saying that pweights are not allowed. Same > problem happens if I specify my weight variable as iw or fw=int(), and also > even when I specify no weight at all. I still get the same error that > iweight or whatever weight I am using is not allowed in this composite > estimation. (aw is not allowed with bs4rw). Perhaps the error messages are > coming from bs4rw? qreg would run okay with any of these weights. > > Any thoughts or suggestions on this problem or on generally doing a BRR on > qreg are greatly appreciated. -pweight- and -iweight- support were added to -qreg- in Stata 13. Unfortunately -bs4rw- is calling all commands under version 9 control. I will try to fix this soon. In the mean time, Mustafa can call -qreg- under explicit version 13 control by typing the following: . bs4rw, rw(brrrwt*): version 13: qreg $depvar $demo if subpop==1 [pw=perwt5], q(.75) If Mustafa is confident that BRR is appropriate, the stern warning/error can be overwritten by specifying "_b" for the <exp_list>: . svy brr _b : qreg $depvar $demo if subpop==1, q(.75) --Jeff jpitblado@stata.com * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/