Determining hay preference in the presence of competition
by
Murray Hannah
Natural Resources & Environment, Victoria
Coauthors: John Reynolds, Gavin Kearney, Stephanie Knott, Dawn Dalley

A project to, "measure the preference of hays in lactating dairy cows", was commissioned at Natural Resources & Environment, Ellinbank, Victoria. The aim was to specify a relative palatability for each hay. To achieve this, a "taste panel' of dairy cows was employed. Hays were presented in pairs to individual animals and the quantity of each hay consumed measured.

An experimental design was constructed to allow for individual animal effects, environmental effects associated with the session, and side (left or right) on which each hay was offered. GenStat Design was of some assistance in achieving good balance of hay treatments over these factors in conjunction with pairwise concurrence, subject to trial size constraints.

The taste-panel design, explicitly pits one hay against another in pairwise comparisons. It is crucial that competition between hays be characterised in the analysis. One approach is to reduce the data to binary outcomes, specifying "preferred" or "not preferred" according to which of the two hays offered in each pair had the greater amount consumed. The Bradley-Terry model (Bradley & Terry, 1952) could then be used to estimate preference "probability" parameters, defining the "palatabilities" of hays. This binary approach, however, discards quantitative information. Furthermore, its interpretation is a little abstract. An alternative, more direct approach, exploiting the quantitative nature of the data, is presented. The method bears some resemblance to a method by Lee (1999) used in the modelling of sporting competition data.

Date received: September 5, 2002

Copyright © 2002 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # cajn-28.