III. Research Design
(i) Research sample
Based on the previous research on the evaluation of online learning behaviour, learning performance model of online courses, and data model of learning analyses, and drawn upon the original data from the learning platform logs in the 2015 autumn term of the Open University of China, data are collected and extracted from course information, student information, teacher information, learning process behaviour and teaching process behaviour etc. Of these, the behaviour data related to learning process were classified according to the students’ learning actions and categories mainly completed online. There are 14 data items across seven categories, including online learning time, frequency of various behaviours, resource utilisation, forum posting, assignments completed, tests completed, and online learning results. The behaviour data for the teachers is mainly collected from three major online teaching behaviours — online tutorial engagement, tutorial methods, and course development — and includes 14 data items in six categories: frequency of various behaviours, online teaching time, forum posting, assignments assessed, internal memos, and course resource renewal. These data items are based on the online course resource category, learning tasks to be completed, and major teaching and support services, including principal learning behaviours of learners’ online learning, learning process and teachers’ principal teaching and tutorial behaviours. Table 1 includes the data items and their specific descriptions. Over the course of data arrangement and collection, incomplete and abnormal behaviour data logs are excluded.
(ii) Research methods
Using applied data statistics and visualisation, cluster analysis, correlation analysis, association rules, and other data analysis and mining techniques, the learning behaviour data from the OUC 2015 autumn term was analysed in order to understand the online learning behaviours apparent in the OUC’s online courses, to mine the relations between and among the behaviour data, and to establish the factors influencing students’ learning behaviours in the OUC’s MPOCs.
An understanding of course development and teaching was gained by tracking the course teaching process. Further analyses of the teaching process were made in order to mine learning behaviour features and to improve course design in combination with quantitative analysis.
A number of tutors and students were interviewed in order to supplement and revise the research results. Explanations of some of the learning behaviours were made by integrating actual teaching experience and practice, and relevant suggestions for instruction and learner support were made.