Bayesian item response modeling

Item response data is analyzed through item response theory, a field also known as modern test theory. Research in this area has grown considerably, yet most of the standard literature relies on frequentist techniques for estimating model parameters and checking model fit. Bayesian methods offer considerable promise for refining how these statistical models are built.

Two qualities make the Bayesian framework appealing for this work. It allows analysts to bring in information that goes beyond the observed responses, and it states clearly how such additional knowledge should enter the analysis. It also relies on simulation-based estimation, which can accommodate many kinds of prior distributions and data-generating models.

The author wrote this book to introduce Bayesian approaches to modeling and analyzing item response data. It is intended as a counterpart to popular item response theory texts by Baker and Kim (2004), De Boeck and Wilson (2004), Hambleton and Swaminathan (1985), and van der Linden and Hambleton (1997), which focus mainly or entirely on frequentist methods. The book shows the usefulness of the Bayesian approach by discussing and applying a range of Bayesian item response models.