2012_12_I/ITSEC - Use of Evidence-based Strategies to Enhance the Extensibility of Adaptive Tutoring Technologies
Abstract: Evolving technology continues to support increasingly advanced training systems that allow customization and personalization of content to provide instruction tailored for individual learner needs. This paper will address the identification of macro-adaptive instructional strategies for informing a generalized model of pedagogy to be implemented in a domain-agnostic Computer-Based Tutoring System (CBTS) framework. Research indicates that higher-order thinking skills are not acquired through didactic approaches but rather learner interaction with the subject matter (Shute & Psotka, 1996). Consequently, it becomes necessary to research strategies that enhance trainees’ learning within computer-based platforms that allow such interaction to occur. This requires prescriptive pedagogy that tailors interaction and feedback based on trainee traits. Intelligent Tutoring Systems (ITSs) are one such application that monitors user interactions and uses Artificial Intelligence tools and methods to assess trainee performance and apply pedagogical interventions to support learning. Here, pedagogical models are responsible for informing adaptation in response to the knowledge state of users by implementing strategies intended to aid in knowledge/skill acquisition. ITSs continue to be effective instructional tools across multiple domains, yet their wide use is limited by associated development costs and lack of extensibility beyond specifically designed applications. To address these constraints, a framework is under development to provide standardized processes for authoring and applying ITS functionality across multiple training platforms and domains. Macro-adaption focuses on using learner aptitude and trait variables, measured prior to training, to inform the system regarding appropriate instructional strategies for achieving maximal learning outcomes. The intent is to utilize research-supported strategies prescribed for specific learner, knowledge, and domain conditions. These parameters will be used to construct a domain-independent pedagogical model for authoring and implementing macro-adaptive functions based on the learner’s historical characteristics. The result will be a self-executing decision tree used to inform and adapt instructional strategies based on known information about the learner.