Breaking Language Barriers: The Power of Machine Translation in Online Learning

This research explored the potential of machine translation (MT) in enhancing the accessibility and inclusivity of online learning platforms, with Coursera serving as a case study. The study compared the performance of courses translated by humans (HT) to those translated by machines (MT) with a toggle feature allowing access back to the original content. The key metrics used were course completion rates and star ratings. The findings reveal that MT courses with the toggle feature have a 2.3 % higher course completion rate than standalone HT versions. Furthermore, MT courses have a higher likelihood of receiving a 5-star rating (88 ^% ) compared to HT courses (84 ^% ). These results consider potential confounding factors such as course pair fixed effects, product line, and learner tenure. Future research could involve experiments that randomly assign learners to different course versions or explore the potential of human translation with the toggle functionality to the original content. The findings have significant implications for practitioners in education and future research, highlighting the potential of MT in enhancing accessibility and inclusivity in online higher education.