Analysis of choice behaviour via probabilistic ideal point and vector models

The aim of this paper is to add results to the everlasting attempt to find appropriate models for the description and analysis of choice behaviour. As stochastic generalizations within choice models which allow estimation and testing in a straightforward way should be of great interest, probabilistic ideal point and vector approaches are used to handle inter‐ and intra‐individual irregularities in paired comparisons preference data. For choice behaviour analysis maximum likelihood estimates of the model parameters are computed and possibilities of testing various model variants by means of log‐likelihood ratio tests are discussed. The proposed models offer information to decide whether an ideal point or a vector approach is more appropriate for the description of the pairwise choice data. Examples of data sets known from previous research in choice behaviour are used for the comparison and for the demonstration of the merits of the proposed procedures.