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: ci's while adjusting for response rate |
| Date | Fri, 21 Jan 2011 12:03:10 -0500 |
Rich <richgold@ix.netcom.com>: I think you are referring to pp.361-363 in Sampling Techniques which address no responses within a stratum. If it is really a simple random sample, no adjustments are needed when one is (perhaps foolishly) using a normal approx CI and assuming responders and nonresponders are essentially identical. A slightly more conservative approach in the spirit of Cochran's discussion (or Manski's work) would be to compute upper and lower bounds assuming the true proportion among nonresponders is 1 and 0, respectively. On Fri, Jan 21, 2011 at 10:19 AM, <richgold@ix.netcom.com> wrote: > Hi, > > I am away from my office and have no access to texts so am asking for help here > > A client has some survey results based on simple random sampling; unfortunately, the response rate was not as good as hoped for; at any rate, they are requesting CI's for some proportions; I recall that Cochran's book (Cochran, WG (1977?), _Sampling Techniques, 3rd edition_, Wiley) discusses an adjustment to CI for non-response rate; if anyone could supply this, and/or thoughts on the use of this adjustment, I would greatly appreciate it > > Rich * * 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/