GAUSSIAN PAIRED-COMPARISON MODELS

Bayesian paired comparison models based on the Gaussian additive error models for the use of expert judgements are formulated. Some limiting properties of the models are investigated. It is observed that the po sterior variance cannot be made arbitrarily small by increasing the nu mber of experts. The posterior distributions are determined by using t he Gibbs sampler, and numerical examples are discussed. The value of i nformation lost in the paired comparison process is analysed from the Bayesian point of view.