Mapping the Knowledge Structure of Students with Intellectual Disabilities Through SOLO Taxonomy-Based Automated Assessment

Background: Mapping students with intellectual disabilities' (ID) knowledge structures is a critical challenge in inclusive education. Traditional assessments often fail to capture the nuanced cognitive development of this population. This limits teachers’ ability to design differentiated interventions. This study aimed to develop and test a SOLO Taxonomy-based Automated Assessment (SOLO-AA) system tailored for students with mild to moderate intellectual disabilities. It also examines its effectiveness in mapping hierarchical knowledge structures. Method: An exploratory sequential mixed-methods design was used. In the qualitative phase, in-depth interviews with six special education teachers and analysis of lesson plans revealed linguistic limitations, fragmented knowledge, and a reliance on visual and contextual learning. These findings informed the design of a visual-interactive assessment system. In the quantitative phase, SOLO-AA was piloted with 30 students aged 11-15 years. Content validity was evaluated using Aiken’s V. Reliability was assessed with KR-20 and Rasch modeling. Result: Content validity reached a score of 0.89 (highly valid), and KR-20 reliability was 0.82 (highly reliable). Most students were at the unistructural (36%) and multi-structural (30%) levels. Fewer reached the relational level (15%), and only 3% reached the extended abstract level. Rasch analysis showed person reliability of 0.79 and item reliability of 0.91. This indicates instrument stability and appropriate item difficulty for this population. Visual-based items were significantly easier. Tasks requiring concept integration were more challenging. Conclusion: SOLO-AA provides a fine-grained mapping of students’ knowledge structures, moving beyond binary judgments. It helps teachers design differentiated instruction, improves diagnostic precision, and offers a scalable AI-assisted solution for inclusive education. This study links SOLO-AA with adaptive automated assessments and shows its utility in special education. Future research should include diverse ID populations, employ longitudinal designs, and integrate culturally relevant content to deepen contextual meaning.