Portal de Periódicos da CAPES

An Introduction to Nonmetric Multidimensional Scaling

1975; Wiley; Volume: 19; Issue: 2 Linguagem: Inglês

10.2307/2110441

1540-5907

George Rabinowitz,

Organic Food and Agriculture

Nonmetric multidimensional scaling methods are useful for spatially representing the interrelationships among a set of data objects. In this, they are similar to factor analytic methods. The assumptions and procedures associated with these methods are, however, somewhat different from those associated with factor analysis, and are more appropriate to certain political data. In this paper the logic underlying nonmetric multidimensional scaling methods is described, and some guides for using these procedures are offered. Nonmetric multidimensional scaling techniques are among the set of procedures available to investigators interested in spatial representation of political objects. These techniques are useful in illuminating the structure hidden in a complex data matrix, and form an important addition to the factor analytic methods which have been widely used in the discipline. They have achieved considerable popularity in recent years, primarily for three reasons. First, they often yield solutions in a sufficiently low dimensionality to permit a visual examination of the structure. This is an invaluable interpretative aid. Second, they permit the investigation of many matrices which cannot be congenially analyzed using factor analysis. Third, they make only