We discuss two challenging scenarios for frequentist and/or Bayesian inference for categorical data. First, for parameter space regions described by order restrictions, frequentist methods are less straightforward than Bayesian methods, especially for interval estimation. Second, for marginal...

Goodman (1979) proposed a class of diagonals-parameter symmetry models for square contingency tables with ordered categories, A simple version of that model is considered in which the log odds parameters have a linear pattern. The model is also a simple quasi-symmetry model. It fits well when...