ABSTRACT
The merit of nonmetric multidimensional scaling as a technique for charting transport cost surfaces is discussed using empirical data. Palpability and speed are chief advantages of the method. General trends rather than fine detail may be detected from the derived cost surface. Changes in transport rate structures are typically interval and not rank-order in kind as transport rates are tied to physical distance. Nonmetric multidimensional scaling is insensitive to these less-than-rank-order variations. Metric multidimensional scaling is proposed for use in situations in which the data to be scaled are related to an absolute referral such as rail distance.