Item Response Theory

Item response theory (IRT) methods are alternatives to classical test theory (CTT) approaches to scale construction, analysis, and scoring. A differentiating feature ofIRTmodeling is the estimation of a mathematical function that relates individual differences on a continuous latent variable to the propensity to respond to a scale item (e.g., respond correctly to a multiple‐choice aptitude item, or respond in category 3 on a five‐point ordered rating scale). Commonly appliedIRTmodels appropriate for dichotomous and polytomous item response data are described. Assumptions underlying these models and interpretation of indices derived from model parameters are presented. Finally, popular applications ofIRTmodels, such as computerized adaptive testing, are summarized.