This report summarizes transfer of training methodologies, beginning with a discussion that defines simulators, fidelity, and realism, and briefly reviews simulator use in training. Two types of training are identified--training for system operation and training for maintenance (troubleshooting). An outline categorizes current research issues on simulation by training strategy, instructional methods, training for normal operation, maintenance and procedural training, training device design, performance measure, cost effectiveness of training devices, and methodology considerations. The evaluation of training effectiveness of simulators is discussed in terms of the experimental paradigm commonly used, performance measurement, and criticism and modification of the paradigm. Simulator fidelity is examined in detail, including its definition, relationship with training, its measurement, and its components, which include stress, environment, layout, wholeness, dynamics, abstraction, and state variables. Research on simulator training is reviewed in terms of new technology, validation of models, acquisition and decay of training, individual differences, skill level variances, performance measurement, decision aids, and mental models. Conclusions focus on the basic remaining problems of using simulators for training, including transfer effect, fidelity level, and their relation to cost. Eighty-eight references are listed. (LMM)