Skip to main content
ARS Home » Research » Publications at this Location » Publication #140954
Title: THE USE OF MULTIDIMENSIONAL SCALING IN THE INVESTIGATION OF RUMINANT FEED PREFERENCE

Author
item Fisher, Dwight
item Dillard, Anthony
item Burns, Joseph
item BUNTINX, S - UNIVERDIDAD DE MEXICO

Submitted to: Agronomy Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 11/10/2002
Publication Date: 11/10/2002
Citation: N/A

Interpretive Summary:

Technical Abstract: A number of interacting factors complicate investigating animal preferences for feeds. When offered a choice between two feeds an animal may choose only one of the pair or may consume both feeds equally. When offered all possible pairs sequentially for a group of feeds, the collective feeding decisions may reflect one selection criterion or multiple selection criteria. When preference experiments are designed to test all possible pairs of feeds then multidimensional scaling (MDS) provides a statistical method of estimating the number of criteria involved in the decision making process. For many years MDS has been used in fields of study such as psychology but its use in ruminant preference is relatively novel. Resulting criteria are expressed as orthogonal MDS dimensions and may be related to measured variables by correlation or regression. We have utilized this procedure in studies of ruminant feed preference and specifically in tests for variation in preference for afternoon versus morning harvests of hay. Ruminant preference among each pair is expressed as a difference ratio. Ruminants often make selections on the basis of multiple criteria related to variables such as digestibility, soluble carbohydrates, and fiber. In this invited presentation the statistical procedure will be demonstrated and the interpretation of results will be discussed.