EFFECT OF ARTIFICIAL INTELLIGENCE-BASED INSTRUCTION ON SECONDARY SCHOOL STUDENTS’ ACHIEVEMENT IN INTEGRATED SCIENCE IN BENUE STATE, NIGERIA

Abstract

This study examined the effect of Artificial Intelligence-Based Instruction (AIBI) on secondary 
school students’ achievement in Integrated Science in Benue State, Nigeria, addressing persistent 
underachievement linked to conventional, teacher-centred methods. A quasi-experimental pre-test 
post-test design was employed with a multistage sample of 60 students drawn from a population 
of 2,450 across 30 public schools. Data were collected using a researcher-developed Integrated 
Science Achievement Test (ISAT; reliability coefficient = 0.87, KR-20) and analysed with mean, 
standard deviation, and Analysis of Covariance (ANCOVA). Students taught with AIBI 
significantly outperformed those taught via the conventional lecture method, F (1, 57) = 48.325, p 
< 0.05, with a large effect size (η² = 0.46), indicating that AIBI accounts for approximately 46% 
of the variance in achievement outcomes. A significant gender difference also emerged, F (1, 28) 
= 5.612, p < 0.05 (η² = 0.17), favouring female students. These findings demonstrate that AIBI is 
a powerful, adaptive instructional tool capable of improving science learning outcomes and 
reducing gender gaps in resource-constrained classroom settings. The study recommends the 
systematic integration of AIBI into Integrated Science teaching in Benue State secondary schools, 
supported by targeted teacher training and technology infrastructure investment, as a practical 
strategy for improving academic achievement and advancing gender equity in science education.