Adaptive Game-Based Learning

Abstract

Rationale of game-based learning (GBL) has its roots in intrinsic motivation (Malone & Lepper, 1987), play theory (Rieber, 1996), and problem-solving (Jonassen, 1997). Emergent tech allowing for adaptive game design requires designers to attend to the automated pedagogical affordances that support the development of knowledge and respond to play variability. Discussses historical basis of ITS and the role it has/could play in adaptive game-based learning systems. Notes how Vygotsky's ZPD provides a useful framework for pedagogical support for learners. Notes tension between the traditional systematic building of knowledge models of learning and the exploratory environments in GBL and situated learning. Games are, by their nature, adaptive, but learning models often aren't. Among other things, the support that the ITS provides cannot interupt the flow of gameplay. The note the importance of being able to assess the knowledge of a student in a non-disruptive fashion and cite's Shute's stealth assessment methods using ECD. Proper knowledge assessment can be fed back into the system to set the appropriate level of challenge. Important Scientific Research and Open Questions Because the concept of adaptive GBL is fairly new, many important questions still remain (e.g., Van Eck 2006). First of all, the means for designing and integrating stealth assessments into games using AI engines is a critical area of exploration. Existing ITS and tutoring models rely on dyadic interactions that are not appropriate for game environments, and current dialogue components of ITSs are not sufficiently sensitive to the context of games. The process of structuring AI models for GBL is made more challenging by the need to incorporate ECD, which presents its own design challenges (e.g., we do not yet have established models that map gameplay to specific cognitive constructs and learning behaviors). Additionally, AI models for GBL must be easily adaptable and modifiable if they are to be employed across the full range of educational settings and audiences. It also remains to be seen whether a game can be designed to not only present content accurately in an open-ended problem space but also to engage the learner. In other words, just because it is possible to create an adaptive GBL environment that is fun, pedagogically sound, and adaptive enough to facilitate effective schema acquisition and assimilation does not mean we can achieve this goal. These are just a few questions researchers will need to explore to realize the promise of truly adaptive GBL environments.