An approach is suggested for the incremental use of real-time ATC simulations for concept development and human factors evaluation of automation systems. Emphasis is placed on the characteristics that distinguish research simulators from those used primarily for training. Four general levels of simulator fidelity are identified for two ATC environments of interest: the radar room and the control tower. Fidelity requirements are generated by the specific needs of the particular human factors study to be conducted, ranging from part-task single-controller simulation used for concept demonstration to full-mission simulation of multiple ATC facilities to examine issues of interaction among automation systems. This approach is applicable to smaller simulations performed at an R&D contractor site as well as large-scale system integration studies conducted at a high-fidelity, centralized, simulation facility. It has been applied to the design of simulations of ATCT and TRACON environments that are being used for the evaluation of displays, controls, and procedures for the Airport Surface Traffic Automation (ASTA) and Terminal ATC Automation (TATCA) projects.