Book Chapter: Item Response Theory
Published in International Encyclopedia of Education (Fourth Edition)., 2022
Recommended citation: Chapter Ackerman, A., Ma, Y., Ma, M., Pacico, J. C., Wang, Y., Xu, G., Ye, T.,Zhang, J., Zheng, M. (2022). Item Response Theory. In Editor R. J. Tierney, F. Rizvi, K. Ercikan. (Ed.) International Encyclopedia of Education (Fourth Edition, 72-85). Elsevier. https://doi.org/10.1016/B978-0-12-818630-5.10010-7
Abstract This article provides a broad overview of item response theory (IRT). It begins with a brief historical background of the motivation and theoretical development of IRT. The goal of IRT is to mathematically describe the interaction between an examinee and an item. This interaction is denoted in terms of probabilistic models which account for characteristics of an item (e.g., difficulty and discrimination) and the latent ability of the examinee. Several types of models are discussed including the unidimensional 1-, 2- and 3PL models for dichotomous response items, the partial credit, and the generalized graded unfolding model polytomous response items. There is also a brief discussion of multidimensional models for tests that require more than one latent ability. The article concludes with a review of algorithms and methodology to estimate the parameters of the models.