Adaptive Literacy Assessment in Education

Abstract

This systematic review synthesises current evidence on adaptive testing models used in literacy assessment within educational frameworks. The review was motivated by the limitations of traditional fixed-form literacy assessments, which often fail to accommodate differences in learner ability levels and may yield inefficient, less accurate measurement outcomes. The study examined how Computerised Adaptive Testing (CAT) and related adaptive assessment technologies improve the precision, efficiency, personalisation, and scalability of literacy evaluation through psychometric frameworks such as Item Response Theory (IRT) and Artificial Intelligence (AI)-supported adaptive systems. A comprehensive literature search was conducted across five electronic databases: Google Scholar, ERIC, PsycINFO, JSTOR, and PubMed. The search yielded 120 studies, which were screened using predefined inclusion and exclusion criteria in accordance with the PRISMA guidelines. After title, abstract, and full-text screening, 10 eligible studies published between 2018 and 2025 were selected for analysis. Findings revealed that adaptive literacy assessment models significantly reduce testing time while improving measurement accuracy and learner engagement. The reviewed studies also highlighted the importance of robust item calibration, algorithm design, AI integration, and psychometric validity in enhancing adaptive assessment systems. Furthermore, evidence from both developed and developing countries demonstrated the feasibility of implementing adaptive literacy tools across diverse educational settings. Despite these advantages, challenges remain in infrastructure, ethical AI use, item bank development, and user experience. The review concludes that adaptive testing models have substantial potential to transform literacy assessment by supporting fairer, learner-centred, and data-driven evaluation practices.