Cognitive Structure Analysis-based Self-assessment and Remediation Greatly Improves Educational Performance and Reduces Score Variance in 8th Grade Science Students
Year: 2026
Type: article
Study design: Clinical Trial
Abstract: Low student performance continues to plague educational systems around the world. The US, for example, has invested billions of dollars in educational reform over the past 50 years in the hope of dramatically raising student achievement but with little to show for this investment. Our previous research shows that Cognitive Structure Analysis-based assessment greatly improves student performance and can bridge the racial achievement gap between Black/Hispanic and White/Asian students. This study investigated whether artificial intelligence (AI)-generated Cognitive Structure Analysis (CSA) self-assessment templates could improve academic achievement and reduce variability in student performance in eighth-grade science. Eighteen students enrolled in a physical science class at a Falcon Academy, a private school in Virginia, were randomly assigned to either a self-assessment condition (n = 9) or a no-self-assessment control condition (n = 9). All students received identical classroom instruction on waves and sound. Students in the experimental condition completed an AI-generated self-assessment template as homework, identifying gaps in their factual, procedural, strategic, and rationale-based knowledge and reviewing instructional materials to remediate those gaps.
Authors: Janat Bajwa, John Leddo
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W7223463363
