The authors have recently developed a microworld, a simplified process control simulator, to simulate a nuclear power plant. The microworld provides an environment that can be readily manipulated to gather data using a range of participants, from students to fully qualified operators. Because the microworld represents a simplified domain, it is possible to have more precise experimental control compared with the complex and confounding environment afforded by a full-scope simulator. In this paper, we discuss collecting human reliability data from a microworld. We review the generalizability of human error data from the microworld compared to other data sources like full-scope simulator studies and compare advantages and disadvantages of microworld simulator studies to support human reliability data collection needs.
Human Reliability Studies With Microworld Simulators
