Effects of an 80-10-10 Multilingual CTCA-Harlybot Model on Achievement and Critical Thinking in Mobile and Adaptive Systems

Abstract

Purpose: This study compared three instructional approaches for teaching a Mobile and Adaptive Systems course: traditional lecture, Harlybot-supported Culturo-Techno-Contextual Approach (CTCA) in English, and Harlybot-supported CTCA using an 80-10-10 English–Nigerian Pidgin–Yoruba code-switching model. It examined their effects on students' achievement (retention) and critical-thinking skills. Methodology: A pre-test–post-test quasi-experimental design involving 75 second-year undergraduates was adopted. Cognitive Task Analysis identified the OSI Model as the most difficult concept, informing the intervention. Quantitative data were analysed using MANCOVA with pre-test scores as covariates, while think-aloud protocols and focus groups provided qualitative evidence. Results: No statistically significant differences were found across the three approaches (p > .05). However, the multilingual CTCA group consistently achieved the highest adjusted mean scores for achievement and critical thinking, with moderate-to-large effect sizes (η² = .09–.14). Qualitative findings showed that strategic, culturally grounded code-switching reduced extraneous cognitive load and enhanced meaningful learning. Novelty and Contributions: The study extends Okebukola's Eco-Techno-Cultural Theory by identifying an optimal indigenous-language threshold (≤20%) for teaching complex computing concepts and demonstrating the effectiveness of integrating AI-supported instruction, CTCA, and multilingual pedagogy. Practical and Social Implications: The findings suggest that limited, purposeful AI-supported code-switching can improve learning while preserving academic rigour, thereby providing a culturally responsive approach to STEM education in multilingual African universities.