A derivation of the Polytomous Rasch model based on the most probable distribution method.

Boltzmann's most probable distribution method is applied to derive the Polytomous Rasch model as the distribution accounting for the maximum number of possible outcomes in a test while introducing latent traits, item characteristics, and thresholds as constraints to the system. Affinities and similarities of the present result with other derivations of the model are discussed in light of the conceptual frameworks of statistical physics and of the principle of maximum entropy.