FAIR-ITEMS: Cognitive Item Models for Automatic Item Generation and Diagnostic Feedback in Large-Scale Assessment

Summarise with VIRAT
Authors: Caroline Hornung, Philipp Sonnleitner
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This project investigates whether cognitive item models — structured blueprints that define which skills an item requires and what features drive its difficulty — can serve a dual purpose in large-scale educational assessment: providing the foundation for template-based automatic item generation (AIG) and directly structuring diagnostic feedback based on Cognitive Diagnostic Models (CDMs). The project was conducted within Luxembourg's national school monitoring program (ÉpStan) using image-based mathematics items across Grades 1, 3, and 5, a context chosen to minimize construct-irrelevant language variance in a highly multilingual student population. Three papers document the approach and its empirical validation: Establishing cognitive item models for AIG, including a novel fairness concept at the item-component level (Differential Radical Functioning) Leveraging cognitive models for CDM-based diagnostic feedback in early numeracy assessment Evaluating hierarchical diagnostic classification models for enhanced student diagnostics in large-scale assessment The results suggest that specifying cognitive models upfront — rather than retrofitting them after test development — pays off across the full assessment pipeline, from item generation through psychometric modeling to instructional feedback.

Fields of Study & Subjects

Psychometric Methodologies and Testing Educational Technology and Assessment Educational Assessment and Pedagogy

Publication details

Journal / source
OSF Preprints (OSF Preprints)
Publication date
20 March 2026
Language
EN
Document type
Preprint