Higher academic performance in an Asian University : replacing traditional lecturing with blended learning

• Main Author: Jones, Kevin Anthony
• Other Authors: Khoo Soo Guan, Christopher
• Format: Theses and Dissertations
• Language: English
• Published: 2018
• Subjects: DRNTU::Library and information science::General
• Online Access: https://hdl.handle.net/10356/88942; http://hdl.handle.net/10220/46012
• Institution: Nanyang Technological University

A young technological university located in Southeast Asia is experiencing problems with low attendance to varying degrees in face-to-face learning sessions, except examinations and labs, and dilatoriness in formative assignments. The result is lackluster and even below-potential academic performance of the students. This is not a unique problem; other universities globally have been experiencing this phenomenon. Although it could be due to several things, such as fragile curriculum and ineffective teaching, in the majority of cases, the other universities have altered or replaced their existing instructional designs. Taking a cue from these global efforts to turn around the diminishing performance of current students, a research project is undertaken to determine the effect of changing the instructional design from traditional lecturing to blended learning on the academic performance on students. The research focuses on two software engineering courses offered to students in different degree programs by the Computer Engineering School in the university. The duration of the research is five years, the first year for measuring the academic performance of students taking the course in traditional lecturing, and the next four years for students in blended learning. The experiment adheres to design-based research methodology, where the design of the blended learning research platform complies with Gagné’s “nine events of instruction” (nine EoI; also known as “nine steps”) and Chickering and Gamson’s “seven principles of good practice”. The blended learning is a mix of eLearning and face-to-face. The eLearning involves motivating videos, prior knowledge assessments, digitized reading materials, learning assessments with feedback, practice exercises with feedback, and surveys on the completed weeks’ learning experiences. The face-to-face comprises problem solving, study groups, presentations, formative assessments, and formative peer evaluations; importantly, traditional lecturing is totally eliminated. Furthermore, the use of educational technology in blended learning, though more than in traditional lecturing, is kept to a reasonable level that should not be an obstacle for other teachers wishing to convert their own courses to blended learning. A two-tier composite model relates formal learning and instruction theories with data processing. The first-tier is a generalized Learning Model based on Prosser and Trigwell (1999), and the second-tier is a Data Model designating the unique variables valuated in the experiment. The Learning Model specifies the elements in learning, primarily learning outcomes, prior experience, and learning approaches. Data Model elements trace “back” to the elements in the Learning Model, and “forward” to the experimental variables. The dependent variable comprises the six categories of marks, and the independent variables are instructional design, nationality, school affiliation, five categories of attendance, team leader, and learning style. The necessary rigour, repeatability, and relevance for the research are handled in several ways. For the entire 11 semesters, there is only one and the same teacher for both courses who follows strict protocols to increase data quality or goodness, measurement of academic performance strictly adheres to Bloom’s taxonomy, the curricula of both courses are locked, and the experiment is live in a real-life classroom setting. To the question of student variance across the eight blended learning semesters, the researcher audit of the integrity of the mark data in a 25% sample of those semesters, confirmed no spurious or inconsistent variance. Several important findings are made. The mean of academic performance achieved in blended learning is higher (statistically significant) than that in traditional lecturing; furthermore, traditional lecturing can be eliminated from higher education without diminishing the learning. Attendance is increased significantly, and appears to be a very effective deep learning approach. As a tool for consistent and transparent adjudication of academic performance, Bloom’s taxonomy is pre-eminent. Student critique on the teacher does decline with the operation of the new instructional design, but not enough to be detrimental to his/her employ. Students from China and India are the only members of the cohorts to experience a drop in academic performance in blended learning. A brief exploration of select global indices suggests this unexpected finding may be due to the ‘digital divide’ existing in those nations. Suggestions for further research in the operation of blended learning include: converting more courses to blended learning, testing the learning styles prediction mechanism, studying the relationship of attendance and academic performance, and analysing prior experience of students with lowered performance in blended learning.