Changes in version 1.4.2
o Modified data preparation: unobserved categories are now retained in the contingency table and the weight matrix on the actual category instead of only its ordering. Weighted analyses on data with unobserved categories (limited to linear, quadratic, interval, and ratio weights) or with zero as the minimum category (limited to interval and ratio weights) previously returned erroneous estimates (thanks again for noticing Roberto).
o To facilitate quality control, the analyzed continguency table and weight matrix are now returned.
Changes in version 1.4.1
o Fixed kra which returned faulty point estimates due to changes in 1.4.0 (Thanks to Roberto Walter)
Changes in version 1.4.0
o Cleaned up the code.
o Fixed maximum kappa which previously returned faulty values for weighted analyses.
o Fixed so that ablines for grand mean, observation mean, and subject mean are computed also with missing data.
o plot, print, and summary set as internal.
Changes in version 1.3.1
o sem based on standard deviation fixed so that it is computed from variance instead of standard deviation. Option to compute sem from the square root of the ANOVA error variance and the consecutive pairwise difference added.
o Plot layout reset to default after use.
Changes in version 1.3.0
o icc and sem added.
o Plot function added for rel objects.
o Ratio weight added to gAC.
o Functions renamed (sorry).
Changes in version 1.2.0
o Numeric data matrices no longer required.
o Incomplete cases omitted listwise for bS, cKap, gAC, and sPi.
o In case of more than two raters, cKap and sPi now generalises to Conger's kappa and Fleiss' kappa, respectively.
o cKap standard error now derived via Fleiss et als., formula from 1969 by default.
o Maximum possible kappa (kmax) and proportion of kappa/kmax added to cKap.
Changes in version 1.1.0
o krA point estimate defined as zero when observed disagreement equals expected disagreement.
o Default number of bootstrap replicates for krA set to zero.
o Functions renamed.