Transfer of training for aerospace operations: How to measure, validate, and improve it

It has been a commonly accepted practice to train pilots and astronauts in expensive, extremely sophisticated, high fidelity simulators, with as much of the real-world feel and response as possible. High fidelity and high validity have often been assumed to be inextricably interwoven, although this assumption may not be warranted. The Project Mercury rate-damping task on the Naval Air Warfare Center's Human Centrifuge Dynamic Flight Simulator, the shuttle landing task on the NASA-ARC Vertical Motion Simulator, and the almost complete acceptance by the airline industry of full-up Boeing 767 flight simulators, are just a few examples of this approach. For obvious reasons, the classical models of transfer of training have never been adequately evaluated in aerospace operations, and there have been few, if any, scientifically valid replacements for the classical models. This paper reviews some of the earlier work involving transfer of training in aerospace operations, and discusses some of the methods by which appropriate criteria for assessing the validity of training may be established.