The Use of Multidimensional Scaling to Determine Principal Resource Axes

Abstract
A multidimensional-scaling technique, called INDSCAL, can uncover consistent dimensions that influence animals in selecting from among an array of choices. The necessary input data are measures of the extent to which animals perceive all possible pairs of choices as similar or different. A demonstration indicates that such an analysis of similarities may succeed in identifying multiple environmental dimensions, but analyses of ranked preferences may be uninterpretable.