Assessing the discriminating power of item and test scores in the linear factor-analysis model

Abstract

Model-based attempts to rigorously study the broad and imprecise concept of 'discriminating power' are scarce, and generally limited to nonlinear models for binary responses. This paper proposes a comprehensive framework for assessing the discriminating power of item and test scores which are analyzed or obtained using Spearman’s factor-analytic model. The proposed framework is organized on the basis of three criteria: (a) type of score, (b) range of discrimination, and (c) conceptualization and aspect that are measured. Within this framework, the functioning and interpretation of 16 measures, of which 6 appear to be new, are discussed, and the relations between them are established. The usefulness of the proposal in psychometric FA applications is illustrated by means of an empirical example.