An examination is made concerning the utility and design of studies comparing nonmetric multidimensional scaling algorithms and their initial configurations, as well as the agreement between the results of such studies. Various practical details of nonmetric scaling are also considered. (Author/JKS)

Several nonmetric multidimensional scaling random ranking studies are discussed in response to the preceding article (TM 503 490). The choice of a starting configuration is discussed and the use of principal component analysis in obtaining such a configuration is recommended over a randomly chosen one. (JKS)

Correspondence analysis can provide spatial or clustering representations by assigning spatial coordinates minimizing the distance between individuals linked by a sociometric relationship. These scales may then be used to identify individuals' locations in a multidimensional representation of a group's structure or to reorder the rows and columns…

Six major problems confronting attempts to use nonmetric multidimensional scaling to represent structures underlying similarity data are identified and the author's prospects for over-coming each of these problems are presented. (RC)

Concerned with another form of analysis of m sets of matrices, the Procrustes idea is generalized so that all m sets are simultaneously translated, rotated, reflected and scaled so that a goodness of fit criterion is optimised. A computational technique is given, results of which can be summarized in analysis of variance form. (RC)

Geometric approaches to representing interrelations among tests and items are compared with an additive tree model (ATM), using 2,644 examinees and 2 other data sets. The ATM's close fit to the data and its coherence of presentation indicate that it is the best means of representing tests and items. (TJH)