The Exploration of Online Engagement Data in LMS as Predictors to E-Learning Outcomes

Abstract

In this study, it is proposed an approach to utilize the students’ online engagement data, in terms of the “counts”, collected by the LMS. Data about 364 students who learned online throughout a semester was analyzed. Due to the skewed and peaked distribution, the negative binomial regression was applied to the data analysis. The test scores and time spent in e-learning produce the significant effects on the log of the counts of the LMS login, the counts of course studying, as well as the counts of the e-pages read. It was shown that using the count outcome variables can form the relationships with the predictors in a linear model.