The dynamical analysis of log data within educational games

Abstract

Games and game-based environments frequently provide users multiple trajectories and paths. Thus, users often have to make decisions about how to interact and behave during the learning task. These decisions are often captured through the use of log data, which can provide a wealth of information concerning students’ choices, agency, and performance while engaged within a game-based system. However, to analyze these changing data sets, researchers need to use methodologies that focus on quantifying fine-grained patterns as they emerge across time. In this chapter, we will consider how dynamical analysis techniques offer researchers a unique means of visualizing and characterizing nuanced decision and behavior patterns that emerge from students’ log data within game-based environments. Specifically, we focus on how three distinct types of dynamical methodologies, Random Walks, Entropy analysis, and Hurst exponents, have been used within the game-based system iSTART-2 as a form of stealth assessment. These dynamical techniques provide researchers a means of unobtrusively assessing how students behave and learn within game-based environments.