Automating Performance Assessment and Diagnosis in Scenario-Based Training

Abstract

Accurate and timely performance assessment is a critical factor in maximizing the training value achieved from a scenario-based training process. It is equally important for assessment to be diagnostic of the behavioral and cognitive factors that affect performance. This paper describes an approach to providing automated performance assessment and diagnosis information to shipboard instructors responsible for training tactical decision making teams. A trainee operator's output actions, derived from keystrokes, is compared to performance criteria and expected actions generated by an expert operator model. Deviations from expected performance provide the input to assessment and diagnostic processes. In keeping with the vision of the Shipboard Instructor Training and Support (SITS) program, the approach is designed to support the shipboard instructor's functions and expertise at various stages in the training cycle.