This paper introduces MatriKS, a new computerized tool for the assessment of fluid intelligence based on Raven-like matrices. Based on knowledge structure theory (KST), a mathematical framework initially designed for efficient assessment and personalized learning, MatriKS is the first large-scale application of KST to fluid intelligence assessment. The validation results for MatriKS, suitable for Italian individuals aged 4 to 11 ( N=568 ), are presented. A multi-method approach incorporating classical test theory (CTT), item response theory (IRT), and KST was employed. Each of the three approaches, with its own assumptions and models, highlights structural properties of the data that are not captured by the other two. Nevertheless, the three approaches provide an acceptable modeling of the data supporting the adequate functioning of MatriKS. The study concludes by exploring the methodological and practical benefits of using KST for constructing tests and estimating individual cognitive profiles.
De Chiusole, D., Epifania, O.M., Anselmi, P., Brancaccio, A., Mazzoni, N., Spinoso, M., et al. (2026). Multi-method validation of the new computerized test of fluid intelligence MatriKS. BEHAVIOR RESEARCH METHODS, 58(7), 1-25 [10.3758/s13428-026-03049-2].
Multi-method validation of the new computerized test of fluid intelligence MatriKS
Spinoso, Matilde;Orsoni, Matteo;Giovagnoli, Sara;Benassi, Mariagrazia;Balboni, Giulia
;
2026
Abstract
This paper introduces MatriKS, a new computerized tool for the assessment of fluid intelligence based on Raven-like matrices. Based on knowledge structure theory (KST), a mathematical framework initially designed for efficient assessment and personalized learning, MatriKS is the first large-scale application of KST to fluid intelligence assessment. The validation results for MatriKS, suitable for Italian individuals aged 4 to 11 ( N=568 ), are presented. A multi-method approach incorporating classical test theory (CTT), item response theory (IRT), and KST was employed. Each of the three approaches, with its own assumptions and models, highlights structural properties of the data that are not captured by the other two. Nevertheless, the three approaches provide an acceptable modeling of the data supporting the adequate functioning of MatriKS. The study concludes by exploring the methodological and practical benefits of using KST for constructing tests and estimating individual cognitive profiles.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
