Document Type : Research Paper
Authors
1 student
2 atu
Abstract
The present study aimed to develop and validate a diagnostic mathematics assessment designed to accurately identify learning difficulties among sixth-grade students, addressing the limitations of traditional evaluation methods. Adopting a developmental–applied approach with an exploratory–explanatory mixed-methods design, data were collected from 620 sixth-grade students in Mashhad through multi-stage cluster sampling.
In the qualitative phase, the cognitive processes underlying mathematical problem-solving were examined using the revised Bloom’s taxonomy, which informed the construction of a Q-matrix and the development of an initial item bank. In the quantitative phase, psychometric properties of the test were evaluated within the frameworks of classical test theory, multidimensional item response theory, and cognitive diagnostic models, including DINA, DINO, and GDINA. Results indicated that the test items demonstrated strong psychometric quality, and the GDINA model most accurately represented the cognitive structure of the assessment.
Additionally, computer-adaptive testing simulations revealed high diagnostic precision and efficiency, even in shortened test versions. Overall, findings suggest that Arta 6 is a valid, reliable, and diagnostically powerful tool that can be effectively applied for early screening and identification of mathematics learning difficulties, providing a sound basis for targeted and evidence-based educational interventions.
Keywords
)