Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Nick Cox <njcoxstata@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Standard deviation with pweight or fweight |
Date | Tue, 27 Mar 2012 01:33:35 +0100 |
You don't say what command you are using. It sounds as if you want aweights. Your data are on schools, not students as such. Nick On Tue, Mar 27, 2012 at 1:29 AM, Alexandra Gecker <agecker@princeton.edu> wrote: > I have a dataset in which each observation is a school i in a year t. > Each observation includes a dependent variable "y_it" and several > covariates "x_it" as well as information on the total school > enrollment. In essence, I want to regress y_it on x1_it, x2_it, x3_it > -- but I want to do so weighting each school by the enrollment. So at > first I thought I would use [fweight=school], but that created a huge > number of observations and therefore very "over-confident" p-values. > > So now I have two questions: > (1) if I use fweight, can I ensure correct standard deviations and > p-values by using the option vce(cluster school)? and (2) if not, > could I use pweight=p_it, where p_it is just the (enrollment in school > i in year t)/(total enrollment across all schools in year t)? > > Thank you so much for your consideration. > > Best, > Alex > * > * 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/ * * 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/