ABSTRACT
The ability to practice skills in simulated environments has long been a central feature of job-related training, particularly in defense and aviation. For example, flight simulators have been used to train pilots for over half a century (Caro, 1988), and the simulation-based training strategy they represent has been expanded to many other contexts, such as driving training (Decina, Gish, Staplin & Kirchner, 1996), business training (Feifer & Bell, 1991), and recently even medical training (Eliot, Williams, & Woolf, 1996). Over time, the underlying simulation technology has changed from electro-mechanical to electronic and digital, resulting in a much smaller physical (and fiscal) ‘footprint’ for the simulated worlds as well as in their greater flexibility and modifiability. In recent years, intelligent tutoring systems have built on the simulation-based training paradigm by adding automated behavioral and cognitive diagnosis to the simulation-based practice environments (Langley, Wogulis & Ohlsson, 1990; Self, 1993), and by using the diagnostic assessments to direct machine-generated individualized remediation and instruction (Towne, 1997).
