Selection programs in livestock populations made use of a wide variety of traits. Among them, phenotypic records for some traits are obtained by a subjective evaluation from a set of experts, like sensory, type, carcass or fat score traits. Data from subjective evaluation usually involves a classification under an arbitrary predefined scale. The output of this process can lead to strong departures from the Gaussian distribution. Moreover, different criteria can be achieved for each expert. In this study, we propose a Slaughterhouse Specific Ordered Category Threshold Model, that assumes a specific observable scale for each specialist, and a common subjacent scale. The procedure is applied to SEUROP conformation score data from the Pirenaica Beef Cattle Breed evaluated at 12 different slaughterhouses from the Basque Country and Navarre.