Yanyan Sheng Bücher

Focusing on the innovative integration of general and specific trait dimensions, this book introduces IRT-based models within a Bayesian hierarchical framework. It assesses their performance in various testing scenarios and compares them to traditional IRT models through Bayesian model choice techniques. The findings indicate that these new models provide a more effective representation of test situations. This work is particularly beneficial for professionals in educational and psychological measurement, offering insights into complex IRT model development and statistical parameter estimation.

Modeling the interaction between persons and items for binary response data, item response theory (IRT) has been found useful in a wide variety of applications. Over the past decades, studies have been conducted on the development and application of unidimensional as well as multidimensional IRT models. However, little literature exists on IRT-based models that incorporate one general trait and several specific trait dimensions. This book, therefore, proposes such models in the Bayesian hierarchical framework, assesses their performances in various testing situations and further compares them with the conventional IRT models using Bayesian model choice techniques. Results from the analysis suggest that the proposed models offer a better way to represent the test situations not realized in existing models. The methodology and analysis should shed some light on the development of complex IRT models and the statistical procedures for parameter estimation, and should be especially useful to professionals in educational and psychological measurement, or anyone who may be considering utilizing IRT models for assessing persons' continuous latent traits.