EFFECT OF THE EDUPHYS AI TUTOR ON THE LEARNING ACHIEVEMENT OF UNDERGRADUATE PHYSICS STUDENTS AT FEDERAL UNIVERSITY DUSTINMA, KATSINA STATE, NIGERIA

Abstract

This exploratory quantitative study investigated the effect of the Edu Phys AI Tutor on the learning achievement of undergraduate physics students at Federal University Dutsin-Ma, Katsina State, Nigeria. Physics education in Nigeria faces significant challenges, including abstract concepts, mathematical demands, and resource limitations. The Edu Phys AI Tutor, an AI-powered learning aid, was designed to address these by providing adaptive content, practice problems, and personalized feedback. Utilizing a pre-test/post-test approach with a single cohort of 18 Physics Education students (15 300-level, 3 200-level), achievement was measured using a researcher-developed Physics Achievement Test (PAT). Content validity of the PAT was established through expert review by three university science educators, and its reliability was confirmed using the Kuder-Richardson 20 (KR-20) formula, yielding a coefficient of 0.87. Statistical analysis, including a paired-samples t-test and Cohen's d, was performed using SPSS. The findings revealed a highly statistically significant increase in mean post-test scores (M = 17.44, SD = 1.338) compared to pre-test scores (M = 13.61, SD = 1.577), with a mean difference of 3.833 points (p < .001). The calculated Cohen's d of approximately 1.72 indicates a very large effect size, suggesting substantial improvement for this group. While this study provides strong initial evidence aligning with the literature on AI tutors' benefits, its small sample size limits generalizability and direct comparative claims against traditional methods. This research suggests the EduPhys AI Tutor's significant potential as a supplementary tool in Nigerian physics education, warranting larger-scale investigations.