Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "Shittu, Aminu" <ameen_vet@yahoo.com> |
To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
Subject | Re: st: Kappa and test of significance |
Date | Thu, 3 May 2012 03:24:50 -0700 (PDT) |
Thank you very much Nick! Aminu. ________________________________ From: Nick Cox <njcoxstata@gmail.com> To: statalist@hsphsun2.harvard.edu Sent: Thursday, May 3, 2012 9:34:17 AM Subject: Re: st: Kappa and test of significance The title of your post although not the text alludes to -kappa- but that seems quite wrong here. The technology of -concord- (-findit- for locations) seems a bit better fit but much of it was developed with measured data, not counted data, in mind. Poisson modelling would seem a more natural starting point. A minimal descriptive approach includes plots of A vs B, ln A vs ln B, (ln A - ln B) vs (ln A + ln B)/2. The last two require all counts to be positive. Nick On Thu, May 3, 2012 at 1:56 AM, Aminu Shittu <ameen_vet@yahoo.com> wrote: > I have a data showing number of parasite counts from composite animal fecal samples from 23 farms. Two individuals (A & B) analysed each sample independently and reported different egg counts presented in the data set as vars acomp, bcomp, and totalcomp = acomp + bcomp. I was wondering whether I could test for agreement and associations between the 2 observers. I will appreciate any help on commands to be used, being a non-proficient Stata user. I use Stata IC v. 10. * * 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/