ABSTRACT
The comparison and assessment of the results of these studies are complicated by several factors, the lack of consistency in the definition of fidelity being the most prominent. In particular, older studies and engineering-based approaches (e.g., Allen et ai, 1986; Thomson, 1989) apparently fail to appreciate the fact that it is not the fidelity of the simulator that is the goal of simulation but the efficient transfer of training. Although these goals are certainly related, they are by no means identical (Lintern et ai, 1989). After reviewing the available literature, Korteling et al (1997) proposed the following subdivision of the term fidelity: physical fidelity (simulator behaviour as specified by the mathematical model, including face validity), and psychological or functional fidelity (pertaining to the similarity of trainee behaviour on the simulator and on the real system). In some cases, then, it is clearly suboptimal to strive for high physical fidelity in simulation (Lintern et ai, 1989; Patrick, 1992). The level of physical fidelity needed to achieve functional fidelity is related to the type of task to be trained, proficiency level, the difference between criterion performance and maximum performance (Boer, 1991a, b), and, last but not least, didactic factors.
