Building a Model to Capture and Analyze Streams of Learning Activity

• In Event: Innovative Learning Design in an Information and Communications Technology-Rich Environment: The Ongoing Story of Shanghai

Abstract

Objectives: This study proposes a learning behavior model that focuses on learning activity streams using an effective analytical method to analyze the streams collected. The purpose of the study is to discover approaches to facilitate and innovate learning analytics in an ICT-rich environment.
Perspectives: A good behavior model and its analytical framework significantly help in collecting, analyzing, and understanding learning behaviors. Using streams of learning activities, the diversified learning sources and the persistence of learning process could be considered in learning analysis. The analytical framework aims to reveal the contexts reflected by the relationships of the elements in the behavior system and build appropriate analytical rules, thus enhancing the efficiency of data discovery and mining.
Methods: This study adopts the research route of “analyzing differences – finding clues – seeking solutions.” In an ICT-rich learning environment, student learning behaviors are more flexible, contextual, and sociable. Therefore, these behaviors can be metaphorized as streams, which indicate the multiple connections with each other and dynamically progress with traceable information. Based on the studies of activity streams (focus on social status) and attention streams (focus on user need), this study proposes the new concept of learning activity streams and constructs its model. For the method of analysis and understanding of the learning activity streams, the context awareness model used by Kofod-Petersen and Cassens (2006) is applied.
Preliminary findings: A simple description paradigm for learning behavior is proposed, which is Actor takes Action [on an Object] [with a Tool] (in a Context). The element feed is specially designed to represent different learning sources. Learning analysis means observing and understanding Action-Object-Tool behavior data that can be used to trace back and mine the potential context information. Contexts are divided into five types, namely, environmental, personal, social, task, and spatio-temporal. The mapping matrix of behavior system elements and context types is established, such as Actor-Object-Object, which can reflect task context to present the network of learning contents. The mapping matrix can also reflect the personal task to reveal the cognitive structure of the learner. The manner of using this matrix for education practice is also discussed.
Significance of the study: The model of learning activity streams is a more appropriate learning behavior model to record and understand the learning activities in an ICT-rich learning environment. Further work is needed, particularly in case studies, to verify and optimize the proposed theory.

Authors

• Xiaohua Yu, East China Normal University

• Xiaoqing Gu, East China Normal University