Effects of AI-supported adaptive learning systems on scientific reasoning, conceptual understanding and misconceptions reduction among pre-service teachers in Kwara State, Nigeria

Abstract

investigated the effect of an AI-supported adaptive learning system on pre-service teachers' scientific reasoning, conceptual understanding and misconceptions reduction in Kwara State, Nigeria. A Pretest-posttest con trol group quasi-experimental design was used. The sample of 120 pre-service teachers was drawn from Kwara State College of Education (Technical), Lafiagi, randomly assigned by intact classes into two groups. An experi mental group (n=60) was exposed to the rule-based adaptive learning system with personal scaffolding, feedback and misconception-targeted interventions, whereas the control group (n=60) was given non-adaptive digital learn ing. The Scientific Reasoning Inventory (KR-20 = 0.84) and two-tier Conceptual Understanding and Misconception Diagnostic Test (= 0.81) were the instruments used. ANCOVA analysis of post-test scores, controlling for pre-test scores as the covariate, revealed a statistically significant main effect of instructional group on scientific reasoning, F(1, 117) = 151.92, p < .001, ηp² = .565, and on conceptual understanding, F(1, 117) = 140.61, p < .001, ηp² = .546, with the experimental group demonstrating significantly higher adjusted mean scores than the control group. Be sides, Pearson product-moment correlation showed that interaction time on adaptive learning was significantly and negatively correlated with scientific misconceptions (r = -.42, p<.01). These results suggest that adaptive learning system is a tool that Nigerian teachers can utilize to positively impact pre-service teachers' scientific reasoning and conceptual understanding. This research recommends that adaptive learning systems with personalized scaffolding and diagnostic feedback mechanisms be adopted and incorporated into teacher education curricula.