I have small uncertainty with Cohen’s kappa coefficient and other kappa coefficients like Fleiss’ kappa or Siegel and Castellan's (1988) fixed-marginal multirater kappa.
Is Cohen’s kappa equal to the two others when we have only two annotators?
Hello Jakub Marszałkowski, Just going from memory Armitage and Berry discuss the kappa statistic, but I've misplaced my book at the moment so sorry I can't check and this is from 10 or so years ago when I looked this up to use the kappa in some analysis I was doing for my MPH. I'll have a look for it, but have the problem of information overload too many books, papers [school, work and church clutter] and email so things are rather disorganised at the moment with my filing. Thanks
Because kappa is influenced by nº observations as the measurements assigned to categories, reason should be replaced with advantage by the models of quasi-symmetry.