Capturing and analysing the processes and patterns of learning in collaborative learning environments

Abstract

<p>This paper describes our methodological experiences capturing and analyzing student learning processes and patterns in three cases. Agent-based models and a virtual world were used for learning. Several specific features tie these cases together and distinguish our analysis from other studies in the CSCL domain. First, students interacted in real-time for relatively short periods. Second, they interacted with each other and with interactive software tools that dynamically 'shaped,' and were shaped by, their learning process. Our work builds upon and integrates process analytic approaches of dynamically captured video and computer screen activity and automatic e-learning process analysis techniques. The first two cases identify areas in which analysis 'by hand' of small amounts of data has produced findings of initial interest. The third case discusses the use of an automatic pattern discovery technique based on Hidden Markov Models to begin to apply these methods to larger data sets in CSCL environments.</p>