Written especially for psychometricians, scale developers, and practitioners interested in applications of Bayesian estimation and model checking of item response theory (IRT) models, this book teaches you how to accomplish all of this with the SAS MCMC Procedure. Because of its tutorial structure, Bayesian Analysis of Item Response Theory Models Using SAS will be of immediate practical use to SAS users with some introductory background in IRT models and the Bayesian paradigm.
Working through this book’s examples, you will learn how to write the PROC MCMC programming code to estimate various simple and more complex IRT models, including the choice and specification of prior distributions, specification of the likelihood model, and interpretation of results. Specifically, you will learn PROC MCMC programming code for estimating particular models and ways to interpret results that illustrate convergence diagnostics and inferences for parameters, as well as results that can be used by scale developers—for example, the plotting of item response functions. In addition, you will learn how to compare competing IRT models for an application, as well as evaluate the fit of models with the use of posterior predictive model checking methods.
Numerous programs for conducting these analyses are provided and annotated so that you can easily modify them for your applications.
Working through this book’s examples, you will learn how to write the PROC MCMC programming code to estimate various simple and more complex IRT models, including the choice and specification of prior distributions, specification of the likelihood model, and interpretation of results. Specifically, you will learn PROC MCMC programming code for estimating particular models and ways to interpret results that illustrate convergence diagnostics and inferences for parameters, as well as results that can be used by scale developers—for example, the plotting of item response functions. In addition, you will learn how to compare competing IRT models for an application, as well as evaluate the fit of models with the use of posterior predictive model checking methods.
Numerous programs for conducting these analyses are provided and annotated so that you can easily modify them for your applications.