The prediction of beef carcass composition: A comparison of models

Any attempt to summarise existing knowledge on the prediction of carcass composition without detailed dissection is bedevilled by the variety of methods used by different workers, not least in the analysis and presentation of their results. This paper examines one cause of the apparent confusion: the differences in interpretation arising from the use of different regression models in allowing for the effect of variable live weight or side weight. The part/whole relationships implied in examining the effect of a total weight on the weight of a part are considered, and by applying four regression models to two sets of published data, it is hoped to facilitate reconciliation of the conflicting results reported in the literature.