A Common Framework for Classical and Tree-Based Item Response Models Including Extended Hierarchically Structured Models

Tutz, Gerhard and Draxler, Clemens (30. October 2019): A Common Framework for Classical and Tree-Based Item Response Models Including Extended Hierarchically Structured Models. Department of Statistics: Technical Reports, No.227

Abstract

A common framework is provided that comprises classical ordinal item response models as the cumulative, sequential and adjacent categories models as well as the more recently propagated item response tree models. The obtained taxonomy is based on the role that binary models play as building blocks of the various models. The study of the binary models contained in ordinal latent trait models clarifies the interpretation of item parameters in classical models. The taxonomy for ordinal models also contains a new general class of hierarchically structured models, which can be seen as a generalization of item response tree models. For this class of models estimation methods are developed, which make use of commonly available program packages.

• Item Type: Paper
• Keywords: Ordered responses, latent trait models, item response theory, graded response model, partial credit model, sequential model, Rasch model, item response trees
• Faculties: Mathematics, Computer Science and Statistics > Statistics > Technical Reports
• Subjects: 500 Science > 510 Mathematics
• Language: English