| Summary: | Massive open online courses (MOOCs) are highly beneficial to the public. However, much research demonstrated low completion numbers of MOOC learners. Several factors have been identified as influential factors in the success of MOOC learners. However, a few studies examined the extent to which prior knowledge is associated with completion rates, learning outcomes, and patterns of student engagement. Hence, this paper aims to examine the relationships between prior knowledge, knowledge gains, and engagement patterns of MOOC learners. Specifically, this study used data mining techniques based on the patterns of learning performance. The Kruskal–Wallis test was used to examine the differences in terms of performance among the detected clusters. Process mining was then used to explore the learning process. The results demonstrated that five groups of learners were identified based on the patterns of their performance calculated from the score obtained from pre-tests (prior knowledge) and post-tests (learning gains) for each of the five topics, namely, Dropout, Stable, Progress, Late dropout, and Post attempt groups. Among the five groups, two of them exhibited different dropout behaviours, namely a) those who dropped out after scoring highly on the pre-test in the first topic in the MOOC and b) those who received relatively low scores in both pre- and post-tests on each topic they studied. This offers a novel insight into MOOC research while indicating that dropout may not be always associated with a lack of success. The study also demonstrated differences in the learning strategies that were adopted by different groups of learners. The key implication of this research is that the introduction of pre-tests as a priming strategy in MOOC design can have strong implications for decision-making related to learning and teaching. © Beijing Normal University 2024.
|
|---|