A model to predict students performance based on instructors pattern of metacognitive scaffolding through data mining analysis

Abstract

The purpose of this study is to develop a decision tree model that can predict student’s performance based on the mechanisms of metacognitive scaffolding prompted by the instructor in Facebook discussion. Prior to the development of the decision tree model, the study identified the pattern of dominant mechanism of metacognitive scaffolding (MS) prompted by the instructor in Facebook discussion. Additionally, students’ academic performance was also investigated. 37 postgraduate’s students from the Authoring System course was participated in a pre-experimental, one group pre and post-test research design. The data were mined using WEKA software and calculated based on the frequency of metacognitive scaffolding posted by the instructor in online learning setting which is Facebook group discussion and also students’ scores in the performance test. The decision tree model predicts that students who achieved grade A in their study were prone to receive a combination of guidance that focused on: i) the process of learning, ii) the rationale for each tasks and activities by the instructor, iii) encouragements in terms of relationship and collaboration among participants and iv) supervision through feedbacks by the instructor such as giving response towards students’ comments. The decision tree model also suggests the appropriate mechanisms of instructor’s metacognitive scaffolding; such a technique should be able to contribute to students’ performance in learning.