Abstract

The article contains a description of Guttman's optimal scoring method, which can be applied to a set of questions that make up a unidimensional scale. The method is almost unknown and scarcely used, even if the computation for its implementation is equivalent to that of multiple correspondence analysis. The differences between the optimal scoring method and correspondence analysis are presented, along with three examples. Optimal scoring method means that internal consistency of the scale, inter-item correlation and alpha coefficient are made rise to the maximum. In a process of developing a test or a questionnaire, the method can help explore the unidimensionality of the underlying construct. Two Internet sites from which it is possible to download the free applications to compute optimal scores are also provided.