Changes in 0.x releases of irr:
=======
 0.84: - An error in the function kripp.alpha was corrected. The old function 
         returned an error when raters showed perfect agreement. (Thanks to 
         Steve Mooney for raising this issue). The function kappa2 now 
         allows for sorting factor levels which becomes important when one
         rater does not use the same rating categories as the other one.
         (Thanks to Sarah Barry for pointing this out). 
 0.83: - The function N2.cohen.kappa was included (Thanks to Puspendra Singh
         for providing this function)
 0.82: - An error in the function icc was corrected. The old function returned
         p-values bigger than 1 for very low icc estimates.
         (Thanks to Elmar Kisslinger for raising this issue)
 0.81: - An error in the function kripp.alpha was corrected
 0.80: - The function relInterIntra was included (Thanks to Tore Wentzel-Larsen
         for providing this function)
       - An error in the function N.cohen.kappa was corrected
 0.70: - The following functions from the concord package were included:
         kripp.alpha (Krippendorff's alpha), rater.bias (test of bias 
         between two raters) and stuart.maxwell.mh (Stuart-Maxwell marginal 
         homogeneity test).
         (Thanks to Jim Lemon for providing these functions)
       - The new function bhapkar was added. It calculates the Bhapkar 
         coefficient of concordance for two raters.
 0.63: - Errors in functions kappa2 and maxwell were corrected
 0.62: - Problematic output characters were removed
 0.61: - Calls to deprecated functions were removed
 0.6:  - Corrected computation of the large sample standard error of the
         weighted kappa coefficient for two raters and the Fleiss' Kappa for
         m raters.
         (Thanks to Tatsuo Kagimura for this hint)
       - Corrected output and documentation of the intraclass correlation 
         coefficient's (ICC) estimate of average score ICCs.
         (Thanks to Jonathan Williams for raising this issue)
       - New functions: agree (computation of simple and extended percentage 
         agreement among several raters), maxwell (Maxwell's RE-coefficient
         for binary scales), iota (interrater agreement for multivariate
         observations)
