Bayesian testing of agreement criteria under order constraints

The most popular criterion to measure the overall agreement between two raters is the Cohen's kappa coefficient. This coefficient measures the agreement of two raters who judge about some subjects with a binary nominal rating. In this paper, we consider a unified Bayesian approach for testing some hypotheses about the kappa coefficients under order constraints. This is done for rating of more than two studies with binary response. The Monte Carlo Markov Chain (MCMC) approach is used for the model implementation. The approach is illustrated using some simulation studies. Also, the proposed method is applied for analyzing a real data set.