Abstract
Training is of little value if trainees can only do the exact tasks on which they were trained, in the identical context of training. Rather, the value typically comes from the ability to apply skills and knowledge across novel variation in contexts and tasks. Training in dynamic technical domains can be particularly challenging because the future tasks can rarely be fully anticipated. We hypothesize that generalization in technology domains will be facilitated when principles (such as device models) are taught in addition to operational procedures, and, particularly, when principles and procedures are integrated. We conducted an exploratory study, including method development, using a micro-world with simulated International Space Station Habitat systems. We compared the effects of Integrated versus Component-wise Training Conditions on generalization to varied tasks, quite different from those in training. Exploratory analyses suggested better generalization and transfer in the Integrated Condition.