Learning Gaps in Science Education through AI: Scale Development among Junior High School Students in a Laboratory High School

The integration of Artificial Intelligence (AI) into education offers promising opportunities to address persistent learning gaps in science, particularly in under-resourced secondary schools; however, few validated instruments assess the impact of AI tools on students' learning challenges. This study aimed to develop and validate the Artificial Intelligence Learning Gap (AILG) Scale, which measures disparities in science education related to AI use by capturing students’ experiences and identifying key dimensions of learning gaps. Employing an exploratory sequential mixed-methods design, the research began with interviews and focus groups involving 20 junior high school students, alongside a literature review that informed the creation of a 4-point Likert scale. The instrument was then administered to 120 students for validation through Exploratory Factor Analysis (EFA) and reliability analysis. The final AILG Scale comprises 29 items spanning four dimensions: Engagement with AI Tools, Cognitive Challenges, Motivation and Personalization, and Teaching Practices. These dimensions collectively explain 41.36% of the variance, with Cronbach’s Alpha values ranging from 0.670 to 0.843, indicating acceptable to high reliability. This scale offers a practical, evidence-based tool for diagnosing science learning gaps in AI-enhanced classrooms, supporting targeted interventions, teacher training, and further research, particularly in contexts where educational technology is becoming increasingly integral.