The Grain Size Of Student Models As A Factor In ICAI Effectiveness

Case-based instruction is a major alternative instructional approach widely used in management training and certainly applicable as well to high-level military training dealing with tactical and strategic decision-making. A wide range of instructionalpractices and technological approaches has been advanced under the broad banner of case-based instruction, with a wide range of claims made and/or demonstrated (to varying degrees), about the benefits of such techniques. We intend to synthesize some ofthe recent advances in cognitive science and information technology into an authoring shell and runtime system that support the process of preparing and delivering effective adaptive case-based instruction for a range of complex skills.SHAI will develop a Runtime and Authoring Shell for Case Analysis and Learning (RASCAL). Given an underlying domain formalization, the system will enable domain experts to author both problem-cases intended to serve as context for instruction, andlesson-cases to serve as useful content remindings during exploration of problem cases. Experts will also be able to author curricula and cases, and link the cases to specified curricular objectives. Depending on availability of case materials, we willdevelop an initial demonstration of this system either in the domain of Naval exercises, or in the domain of business strategy. An authoring shell that eases construction of case-based instructional systems and helps insure fielded systems exploit the best known pedagogical technique and implementation technology will lower costs while improving quality and availability for animportant class of computer-based adaptive instruction. Case-based instruction is appropriate in training a wide range of high-level skills, ranging from military command reasoning to business and other forms of professional decision-making.