Enoch O. Oladunmoye, PhD

Item Response Theory as a Modern Framework for Psychological Measurement

Classical Test Theory (CTT) has provided the foundation for much of psychological and educational measurement, but contemporary psychometric research increasingly requires models that describe measurement at the level of individual items and across the latent trait continuum. Item Response Theory (IRT) provides such a framework. Rather than treating a total test score as the primary unit of analysis, IRT models the probability of a particular response as a function of an individual's latent trait level and the characteristics of the item that produced the response. This approach makes it possible to estimate item difficulty or location, item discrimination, guessing where theoretically appropriate, and measurement precision across different regions of the latent continuum. This paper introduces IRT as a modern psychometric framework and situates it in relation to Classical Test Theory concepts that underpin scale development, reliability and validity work. It explains the latent trait concept, the item characteristic curve, the major dichotomous IRT models, item and test information, the standard error of measurement, model assumptions, calibration, scoring and applications in psychological research. Particular attention is given to a persistent misconception, namely that IRT is simply a "better version" of CTT. The two frameworks answer different measurement questions and rest on different assumptions, and a mature measurement practice draws on both. The paper proposes a practical IRT workflow, beginning with dimensionality and local independence diagnostics and proceeding through model selection, item calibration, item fit evaluation, information analysis, score estimation, differential item functioning analysis and interpretation. The framework positions IRT as a foundation for computerised adaptive testing, psychometric scale refinement, measurement fairness and next generation psychological assessment.