| Sumario: | Owing to the major drawback of analyst Louis Guttman scaling technique, i.e. the lack of adequate criteria for the simple placement of non-scale response patterns in the scale, this research paper describes a simple method for the placement of non-scale response patterns in a Guttman scale. The non-scale patterns are those which can be classified uniquely in a scale type by the minimum error criterion. Without using the complex "Modified Latent Distance" technique, types, where the computations involved require eight to twelve hours' work by a skilled computer and computation becomes problematic as the number of items in the scale increases beyond four, this paper adopts a new approach. In this approach of role conflict scale i.e. non scale response patterns are first assigned by the minimum error criterion. If it results in ambiguous assignments then the patterns are assigned to the perfect scale type having the greatest frequency of cases. Of the 216 role conflict cases, only nine are assigned ambiguously by the method suggested in this paper. In seven of the 216 cases, the "distribution of perfect types" placement disagrees with the latent distance placement.
|
|---|