Blended Learning Can Improve Student Grades – Narxoz University's Large Scale Experiment Shows

Abstract

To-date, blended learning research has been based on case studies, descriptive analysis or on small sample quantitative studies. Few large data set studies were conducted, with little control over the experiment conditions. This article presents a large scale blended learning experiment, involving 295 blended learning students. The related econometric analysis compared them to almost 3000 students in control groups. We found that blended learning courses moderately improve student learning outcomes compared for those students who manage to finish the course and take the final exams. For pure online learning the effect was negative though, due to the fact that the new Moodle platform created a more rigorous learning environment. We also found that teacher quality and the students’ financial situation strongly influence student grades, and we validated the Best-In/Best-Out (BIBO) rule for higher education, i.e. that universities that attract the best freshmen will likely also have the best graduates.

Keywords: blended learning, learning outcomes, experiment, econometric evaluation, large data set