Investigating Learners' Interaction Patterns in Synchronous and Asynchronous Online Discussions and Their Relationship with Learning Performance

Learner interactions are necessary for academic success and can be implemented synchronously or asynchronously in online education. Taking an online course from a university in Central China as an example, this study adopted social network analysis and k-means clustering methods to quantify and cluster learners' interaction patterns and further explore their relationship with learning performance. Based on the interaction patterns observed, learners were classified into four clusters: "double active", "double negative", "synchronous active", and "asynchronous active". In addition, learners in the "double negative" cluster had the lowest level of learning performance.