Search

The purpose of this monograph is to show that sufficient achievements in scaling applications have been made to justify serious study of scaling methodologies, particularly multidimensional scaling (MDS) as a tool for geographers. To be useful research, it was felt that the common methodological and technical problems that specialized researchers share with other scholars should be indicated by review of the applications, and that an adequate statement on the mathematics and heuristics of scaling algorithms is necessary. As a review of applications, subroutines in scaling programs are "dissected" in order to understand how certain critical parameters are defined and used. This research work is presented in three parts relating to 1) basic fundamentals of scaling, data requirements, and algorithm constructions and problems; 2) two step-by-step examples of the non-metric section of a multidimensional scaling algorithm; and 3) a review of geographical applications of the approach in a variety of problem areas. The position of this paper is that MDS provides a useful and constructive methodology for examining the problems of preference and choice for researchers in geography. In conclusion, some problems of using MDS are mentioned and its potential uses in geography given. (Author/ND)