An Improvement Over Guttman Scalogram Analysis: A Computer Program for Evaluating Cumulative, Nonparametric Scales of Dichotomous Items

Publisher Website
Google Scholar
Abstract
This article mentions some of the problems associated with the criteria of reproducibility and scalability as they are commonly used in Guttman scalogram analysis to evaluate cumulative, non-parametric scales of dichotomous items. It then describes the advantages of criteria developed by Loevinger (1948) and Mokken (1971) to evaluate such scales. A short SAS computer program is presented which analyzes response patterns elicited by dichotomous scales designed to be cumulative. The program computes Loevinger's H as index of the scalability (cumulativeness) of a set of items as well as Mokken's Hi for each item indicating the degree of scalability of each item with respect to the other items of a given scale.