Explanatory item response models

Autor*innen

• Paul De Boeck

Reihen

• Statistik für Sozial- und Verhaltenswissenschaften

Parameter

• 386 Seiten
• 14 Lesestunden

• Psychologie

Mehr zum Buch

This edited volume offers an integrated introduction to item response models, commonly utilized in psychology, education, and social sciences, through the lens of generalized linear and nonlinear mixed models. This new framework broadens the scope of item response models, highlighting their explanatory potential beyond traditional descriptive applications. The core idea is that item responses can be modeled based on various predictors, which may include characteristics of items, individuals, or their combinations, and can be either observed or latent, continuous or categorical. This approach generates a diverse array of models, encompassing existing item response models and introducing new ones, with particular emphasis on those featuring explanatory predictors, while also addressing descriptive models. The term "item responses" extends beyond conventional test data to include categorical data from repeated observations or longitudinal designs. The book begins with an introductory section followed by chapters detailing models for ordered-category data, multilevel models, differential item functioning, multidimensional models, local item dependency, and mixture models. It includes a chapter on statistical background and software, utilizing a unified notation approach and a consistent dataset for illustration. Computer commands from the SAS package, along with examples from other software, are provided for model estimation. The

• Sprache: Englisch
• Erscheinungsdatum: 2004