Evaluating the effectiveness of the medium of instruction on learning outcomes in technical subjects among Indian undergraduates: A quasi-experimental cross-over study

Abstract

The medium of instruction plays a pivotal role in shaping students’ academic performance, engagement, and learning behavior—particularly in linguistically diverse contexts like India. This study investigates the pedagogical efficacy of bilingual (Hindi-English) versus monolingual (English-only) instruction in technical education, focusing on two core Computer Science subjects: Operating System Concepts and Object-Oriented Programming using C++. A quasi-experimental cross-over design was employed with undergraduate students, integrating pre-tests, four formative assessments, post-tests, structured feedback, and instructor observations. Quantitative results revealed that bilingual instruction led to significantly greater learning gains in abstract topics like Memory Management, as evidenced by paired t-tests and Mixed ANOVA. In contrast, programming concepts such as Constructors and Operator Overloading showed no significant post-test differences between the two modes, suggesting mode-neutral efficacy for syntax-driven content. Qualitative insights indicated enhanced clarity, emotional comfort, and active participation under bilingual instruction, particularly for students with limited English proficiency. Instructors observed improved in-class interaction and fewer post-class queries during bilingual sessions. This study fills a critical gap by offering empirical evidence on instructional language design in STEM education. It supports the development of a flexible bilingual teaching approach that improves student achievement and serves as a step toward expanding multilingual practices in higher education.