Doing smallest space analysis

Smallest space analysis (SSA) is one among the new methods of nonmetric analysis developed by Guttman (1968) in Jerusalem and computerized by Lingoes (1965a; 1965b; 1966a; 1966b; 1966c; 1966d; 1966e; 1967a; 1967b) at Ann Arbor; these methods are recommended for those jobs where the investigator desires a rigorous multivariate analysis under the constraints of no special assumptions. A pleasing related feature of the techniques discussed here is that the results achieved are directly and intuitively interpretable by relatively untutored persons, as well as by the scientist who takes responsibility for his project in its entirety. From the point of view of a user, SSA refers to a set or family of computerized programs designed to accommodate a variety of different conditions, but here the description will be confined to only two of these, SSA-I (for symmetric matrices) and SSA-II (for asymmetric matrices). If observations can be assembled according to the paradigm in Figure 1, and if the number of variables does not exceed computer limitations (about 50 on the IBM-7040, 100 on the IBM-360), then SSA may be used. The researcher wishing to perform an SSA must select some measure which expresses observed relationships among his variables taken in pairs. Pearson's r, Spearman's rho, Goodman and Kruskal's gamma, lambda as