IV. Research Process and Discussion

(i) Features of students’ online learning behaviour

1. The distribution of time spent learning online

Figure 1 is a tendency chart of the days the students spent learning online based on the learning behaviours recorded on the learning platform in the 2015 autumn term. It can be seen that most students learned online for just one day and that only 14 students learned online for more than 40 days. Among the students, the maximum number of online learning days is 59 days. Based on the number of weeks offered for each course (each course is offered over approximately 10 weeks, excluding examination week), the students clearly spend fewer days on online learning, and many of them spend one day or several days on completing the items for their achievements, such as assignments and scored tests. Only a few students maintained long-term online learning. Though much attention has been paid to the rich learning resources and learning activity design of the OUC’s online core courses, in practice the overall online utilisation rate of MPOCs doesn’t make a big difference to the overall online utilisation rate compared with that of MOOCs if there is no follow-up teaching organisation or complete support service system. Students don’t necessarily conscientiously insist on learning through online network courses just because they are formally registered and have paid for admission.

Figure 1 Tendency Chart of Students’ Days Spent Learning Online

2. Comparative analysis of various learning behaviours

The behaviours of the students are of significance in understanding their learning features. The students’ preferences are identified through an analysis of their various behaviours in order to provide evidence for course design and teaching implementation. Figure 2 is a comparison chart of the mean value of four categories of different behaviours (the average value of the behaviours is the average of all the students’ clicks on each course) from the 57 courses on the learning platform for the OUC’s 2015 autumn term. It can be seen that the rarest behaviour is resource browsing, followed by behaviours related to interpersonal interaction. The best click rates are found among other categories of behaviours such as course entries, navigation, and results queries.

Figure 2 Comparative Chart of the Average Value of Various Behaviours

Note: ①The horizontal and vertical coordinates represent course codes and clicks, respectively

To be more specific, scored assignments and tests get the most clicks. Figure 3 and Figure 4 are examples of one public compulsory course and one special core course, respectively. The Figures show the most clicks on the courses’ top ten items. The Figures demonstrate that the interpersonal interaction and human-machine categories of behaviours account for most of the clicks; the top four ranking items are almost all formative assessments (namely assignments), it is the similar case with most of the other courses.

Figure 3 Clicks on Items within OUC Course A

Figure 4 Clicks on Items Within OUC Course B

By integrating course practical teaching and interviews with teachers and students, we established that most online courses have few activities in the interpersonal interaction category, such as assignments and forums, and that the design of course navigation and learning pathways is complicated. With regards to single course items, assignments receive the most clicks, but the overall clicks are less prominent. Furthermore, unclear course learning pathways also lead to repeated clicks of navigation items, which adds to the number of clicks on other categories of behaviour interactions. In the meantime, since degree education students won’t receive a score unless they complete their assignments and tests, they pay more attention to the tests and assignments which can help them obtain scores and less attention to items such as resource browsing.

3. Group distribution of students’ learning behaviours

K-means clustering analysis is used to analyse the students’ online behaviours and their days spent learning online, as shown in Table 2 and Figure 5. Excluding individual numerical outliers, the students’ online learning group is divided into four categories. As seen from Table 2, the total sum of learning behaviours positively correlates with an increase in online days, and Cluster 4 has the most online learning days and behaviours. If the class time is calculated according to 10 weeks in a term, the average length of online learning is 1.5 times a week, and there are 137 different learning behaviours each time. The students with a large number of learning behaviours account for only a small proportion, merely 0.28%. The days spent learning online for Cluster 1 is similar to those of Cluster 4, but the total learning behaviours are less than half of that of Cluster 4; the days spent learning online for Cluster 2 are slightly fewer than those of Cluster 1, and its total learning behaviours are about one third those of Cluster 1; the online learning days and learning behaviours of Cluster 3 are obviously decreasing, and its learning days are about one third of Cluster 2 and its learning behaviours only one fifth of Cluster 2. Figure 5 is a group distribution map of the online course students’ number of days spent learning online and their learning behaviours. The students with one log in or above in an average week account for less than 12% and most of the students log on to learn once in three weeks with an average of three learning behaviours.

Figure 5 Group Distribution of Online Learning Behaviours

Further analysis of the cluster results demonstrates that that the 57 courses of the other three clusters are distributed at random, with the exception of Cluster 4, whose students are all participating in the “OUC Online Teaching Team Pilot Course”. These courses have their own features. In terms of construction, the teaching team includes course leaders, tutors, class tutors, administrators, and technicians. There is an online duty roster, so that the teachers can answer questions raised by the students. In terms of teaching implementation, the role of the team members is to offer students timely tutorials and facilitate their learning. The class tutors, in particular, keep tracking the students’ online learning. They get regular notifications about assignment completion and take initiative to contact students in order to urge them to learn online and to submit assignments.

4. Course interaction

Figures 6 and 8 are descriptive schemes of the relationship between students’ human-machine interactions (tests) and interpersonal interactions (assignments and forum posts) as part of the courses. It can be seen that neither test completion, assignment submission nor forum posts are very satisfactory. The highest test completion rate and assignment submission rate for the courses is still less than 20%. The mean value of forum posts for many courses is 0-1, indicating that some students on the courses do not post any forum posts.

Excluding the 19 courses with no assignment design, the highest assignment submission rate among the rest of the 57 courses is 80.11%. There are nine courses with an assignment submission rate of 50%, accounting for 23.7% of the total. Excluding the six courses with no test design, there are 10 courses with a test completion rate of 50%, accounting for 19.6% of the total. 21.47 posts on the students’ forum represents the highest average number of posts; there are nine courses with an average of five posts by students.

By tracking the teaching activities of several courses with good human-machine and interpersonal interaction, and holding interviews with tutors and students, it can be seen that these courses are successful in both course design and the implementation of the teaching process. For example, some courses are provided with sufficient teaching staff , and some courses are included in the “OUC Online Teaching Team Pilot” project. The students are guaranteed effective and timely tutorial and learner support. Some courses are designed concisely and clearly, and the original flow layouts of the platform are generally maintained, which are well received by the tutors and students. Neither teachers nor students will get lost in the process of teaching and learning. For example, the flow layout is used in the OUC course Organisational Behaviour, and all the resources and activities are presented on the home page. There are no other designs except for learning content on the course page. It is very convenient for tutors to organise teaching or to add resources. Furthermore, it is easy for students to find what they need to learn effectively. Therefore, it ranks on top both in terms of the completion rate for assignments and tests and in terms of forum posts. It is easy to operate and the teachers and students praised its availability in the interviews.

From further observation of the course entries, it is found that downloads are offered for scored tests and assignments to make it easy for the students to finish assignments. The students can download and complete the assignments and scored tests offline. This is a common practice among students in many branches, especially those in less developed areas. Meanwhile, most of the classes have established QQ and WeChat groups, with which to conduct real-time and non real-time interactions. This leads to low forum use rates, accounting for the low levels of human-machine and interpersonal interaction on the platform.


Figure 6 Distribution Map of Test Completion Rates


Figure 7 Distribution Map of Assignment Submission Rates

Figure 8 Distribution Map of Students’ Course Posts