On Exploring Order Effect and Ties in Paired Preferences with Bayesian Analysis

Abstract

The paired comparison technique is used to rank stimuli on the basis of preference data obtained through presenting stimuli to some respondents in the form of pairs and asking them to record their preferences of the stimuli based on sensory evaluations. In this study, the Kuk (1995) model for paired comparisons, which accommodates home-ground advantage (equivalent to the order effect) and draws/ties in the paired comparison experiments, is considered for analysis in the Bayesian framework. Worth as well as tie parameters are estimated for both at home-ground and away-from-home matches. The entire estimation procedure is illustrated using a real dataset