
The Rasch model is designed for categorical data, often collected as examinees' responses to multiple tasks such as cognitive items from psychological tests or from educational assessments. The Rasch model's elegant mathematical form is suitable for extensions that allow for greater flexibility in handling complex samples of examinees and collections of tasks from different domains. In these extensions, the Rasch model is enhanced by additional structural elements that either account for differences between diverse populations or for differences among observed variables.
Research on extending well-known statistical tools like regression, mixture distribution, and hierarchical linear models has led to the adoption of Rasch model features to handle categorical observed variables. We maintain both perspectives in the volume and show how these merged modelsa "Rasch models with a more complex item or population structurea "are derived either from the Rasch model or from a structural model, how they are estimated, and where they are applied.
| samuel butler n a jairazbhoy alan w biermann marc f lieberman thomas bailey aldrich | helga nussbaum allan h goodman walter veltroni naveed sherwani p a caraveo |