F1-AS1-1 - A Unified Approach to Quantitatively Measure the Similarity of Computer Science Units in Australia

3. Research Full Paper
Sameera Jayaratna1 , Timos Sellis1, Caslon Chua1, Mohammed Eunus Ali2
1 Swinburne University of Technology
2 Bangladesh University of Engineering and Technology

Computer Science education is increasingly becoming popular across the globe. According to the Australian Computer Society (ACS), the annual enrolments in Computer Science (CS) / Information and Communications Technology (ICT) related programs at universities in Australia are 50 percent higher than a decade ago. With the increasing demand, most of the higher education providers offer variants of programs in CS/ICT. Currently, around 40 education providers with a total of 340 CS/ICT programs hold current ACS accreditation. With the increased popularity of CS/ICT courses, the ability to measure the similarity/distance between units (e.g Introduction to Programming, Software Engineering, etc.) can aid the decision-making process of education providers during learning path recommendation, evaluating unit exemptions/advanced standings, etc. In the case of evaluating units for exemptions/advanced standings, the evaluator needs to compare two units based on their contents, learning outcomes, etc. This is often done by a university academic in the relevant field. The ability to quantitatively measure the similarity of two units will assist this process which is otherwise likely to be tedious and time-consuming. Also, the process of learning path recommendation involves identifying units that are similar to the units which a given student has completed and other students with a similar learning profile to a given student. The existing learning path recommendation approaches may benefit from the ability to cluster units based on their similarities. Therefore, in this work, we explore how to quantitatively measure the similarity/distance between two CS/ICT units, in an Australian setting. The data for this study is generated as a result of the ACS course accreditation process during which each unit of a course is mapped into knowledge areas. ACS introduces 19 knowledge areas and the process of mapping units to knowledge areas follows Bloom’s Taxonomy. In this study, we utilize data from multiple CS/ICT courses offered by two Australian education providers. We used multiple distance measures to evaluate the similarity of two units based on their coverage of the 19 knowledge areas. Furthermore, we used hierarchical clustering to cluster units based on their similarities to each other. Our results indicate that the similarities can be successfully measured which also can be explained intuitively. Bloom’s Taxonomy is often used in prior research in computing education, including studies on course comparison in areas of education other than CS. Our work is novel in examining a quantitative measure of unit similarity, by introducing a unified approach across education providers that follow the ACS course accreditation process.