学习分析技术应用:寻求数据支持的学习改进方案

Linking Learning Analytics with Instruction Practices: Approach to the Data-enabled Research to Learning Enhancement

Learning analytics is fast growing as one of the leading-edge topics in learning science due to the dramatically accumulated data in education system. The wide spread JCT has resulted in the phenomenon of fl big data fl within various learning technology systems and may have great potential for education researchers to obtain an in-depth understanding of student learning behaviors and performances and to foster data-driven education improvement. One of the biggest challenges is how we can collect， analyze and present the appropriate evidence by integrating researchers， expertise and stakeholders' expectations. The purpose of this paper is to present a series of studies on learning analytics for improving learning outcomes in educational settings. These studies were designed to provide data-enabled learning enhancement with slightly different approaches. The first case was designed to enhance problem-solving outcomes， by a deliberately designed assessment approach. A three-stage approach was proposed to analyze problem-solving behaviors ， abilities ， and performances ， and was applied to assess five hundred and fifty-four students' online problem solving in a primary school. The study reveals four clusters with distinctive problem-solving behaviors， abilities， and performances. The analysis method proposed in this study can also be used to provide interpretation of the problem-solving abilities for each and every student. The second case was designed to enhance online learning by developing a prediction model， from which the instructor can provide timely intervention to facilitate online students' learning behaviors. In this case， a comprehensive model of learning analytics was established， in which the explicit and implicit learning behaviors were both included as the main data source. Clustering analysis and decision tree analysis were applied to reveal students， characteristics and predict their performances. The research result would be of practical value to fully motivate learning and promote learning outcomes. The third case was designed to provide visualized results of learning analytics as a way to promote students， self-efficacy， given the premise that students， beliefs on their academic capabilities play an essential role in their motivation for achievement; as a result， the way that intervention was provided in this study could enhance learning by improving students’ self-efficacy. As shown in the result， this study did hold the hypotheses that visualized learning analysis has an effect on promoting students’ self-efficacy， and thus influence the learning enhancement. By presenting the above research cases dedicated to enhance learning with learning analytics methods and techniques， this paper further discussed the possible ways to link the research of learning analytics with the practical needs of learning enhancement.