Based on learning analytics, this study investigated whether learners’ time management data (the regularity of connecting to the system, total learning time, the number of accessing the system) accumulated on learning management system (LMS) could p...
Based on learning analytics, this study investigated whether learners’ time management data (the regularity of connecting to the system, total learning time, the number of accessing the system) accumulated on learning management system (LMS) could predict their total scores that determined the completion of a course in distance life-long education. It also attempted to determine learning progression points that could predict learners’ course completion significantly and to classify them into three status groups, a risk group, a border group, and a safety group at each significant time point.
This study was conducted with the consent of a typical distance lifelong education center (DLEC) located in Seoul. The subjects of the study were 411 students who had taken five courses offered by the DLEC in the spring semester of 2016. It was assured that they did not take more than one course among the target courses of the study. A log data of the subjects was anonymized and extracted from the LMS of the DLEC and the regularity of connecting to the system, total learning time, and the number of accessing the system were calculated by tracing the time to connect and disconnect to the LMS. After refining the data, descriptive statistics of all the research variables―the regularity of connecting to the system, total learning time, the number of accessing the system, and total score―were conducted and the correlations between the variables were also checked. Stepwise multiple regression analysis was conducted to examine the relationship between the variables related to time management and the total scores. In order to decide learning progression points at which total scores can be predicted, multiple regression analysis was conducted each quarterly point of total 16 weeks for a course to predict total scores in terms of time management data accumulated. At each quarterly point that proved to be significant to predict total scores, they were calculated and learners were classified into three status group. The results of the study are as follows.
First, testing the significance of predicting total scores two time management variables the regularity of connecting to the system and the number of accessing the system proved to predict total scores significantly. More regularly students connected to the system, higher total scores they recorded. More frequently they accessed the system, higher total scores they showed. However, total learning time was found to have no significant effect on total scores.
Second, in order to find learning progression points at which total scores can be significantly predicted, students’ total scores were estimated based on the time management data accumulated every four weeks. As a result, the regularity and number of connecting to the system were found to predict total scores significantly from the first quarterly point: 32.4% of the students who did not actually completed the course were categorized into the "risk group" showing the lowest probability of course completion. This means that just four week time management data can fairly predict the students who are at high risk of course incompletion. Therefore, it is concluded that learners’ total score, which determines their course completion, can be predicted from the 4th week after opening a course that runs at 16 week semester base. Thus, it is possible to establish a support plan especially for the learners who are at high risk of course incompletion from the first quarterly point of the semester.