New Perspectives in Science Education
AI-Supported Differentiated and Personalized Instruction in Mechanics: A Quasi-Experimental Study in Moroccan High School Physics and Chemistry Education
Mouna LAATAR, Scientific Research and Pedagogical Innovation Laboratory (ReSIP), Regional Centre for Education and Training Professions, Madinat Al Irfane, Rabat (Morocco)
Amina El Kasri, Research Laboratory in Sciences and Techniques, Ecole Supérieure de l'Education et de la Formation (ESEF) - Ibn Zohr University (UIZ), Agadir (Morocco)
Mohammed ABID, Scientific Research and Pedagogical Innovation Laboratory (ReSIP), Regional Centre for Education and Training Professions, Madinat Al Irfane, Rabat (Morocco)
Abstract
This study investigates the contribution of artificial intelligence (AI) to differentiated and personalized instruction in mechanics within high school physics and chemistry education in Morocco. It examines whether AI-supported instructional strategies can enhance students’ conceptual understanding and achievement while supporting differentiated learning pathways. A quasi-experimental study was conducted with 60 students (12th grade) enrolled in the common core scientific track (French option), comprising a control group (n = 30) receiving conventional teacher-guided differentiated instruction and an experimental group (n = 30) benefiting from an AI-supported instructional design using Google Gemini for learner profiling and task differentiation and MagicSchool AI for formative scaffolding. Students’ performance was assessed through pre- and post-tests covering key mechanics concepts, including forces, motion, inertia, and equilibrium. Data were analyzed using the IBM SPSS Statistics software package. Analyses included descriptive statistics, independent- and paired-samples t-tests.
The findings indicate that, before the intervention, the experimental group scored slightly higher than the control group. Both groups improved between the pre-test and the post-test, although the experimental group achieved a higher raw mean score at post-test. The experimental group's higher post-test performance cannot be attributed to the AI-supported intervention with confidence, although a small effect cannot be excluded. Item-level comparisons showed, on a descriptive basis, that the experimental group performed better on tasks requiring more abstract conceptual reasoning, which might suggest benefits for certain skills; however, these comparisons were exploratory in nature. The results suggest that AI may be explored as a complementary tool for differentiated and personalized teaching in physics and chemistry, while emphasizing the need for larger sample sizes, groups with comparable prior levels, and further controlled studies.
Keywords: Artificial intelligence; differentiated instruction; personalized learning; physics education; mechanics; secondary education.
