Predicting and mitigating human error in manned spaceflight can be the difference between mission success and lost vehicle or crewmember. The National Aeronautics and Space Administration (NASA) has used the Cognitive Reliability Error Analysis Model analysis developed by the nuclear industry during the last 30 years of manned spaceflight to predict human error. Although the analysis has proven to be reliable, it does not take into account operations specific for long duration spaceflight such as crew training and ground support. This article first explains the principles of the Cognitive Reliability Error Analysis Model and how it is used at NASA. Then, the probability for error for an International Space Station ingress procedure is calculated using standard performance shaping factors developed for the nuclear power industry. Lastly, the environmental and operational constraints of space flight are used to develop new performance shaping factors specific to a NASA‐operated spacecraft. Copyright © 2012 John Wiley & Sons, Ltd.

Human reliability analysis is a crucial for manned spaceflight success. Cognitive Reliability Error Analysis Model (CREAM) has been developed and used by the nuclear industry in predicting human error. Previously, the authors have calculated the probability error for an International Space Station ingress procedure using performance shaping factors (PSF). In this paper, the procedural risk under both ideal and common conditions using the new spaceflight specific PSFs is calculated. The risk was found to vary from the risk calculated using standard PSFs and to vary greatly depending on the spacecraft specific conditions. Under ideal conditions, the risk was found to be 1 in 88, but under common conditions, the risk was 1 in 3. Then, the new PSFs were used to analyze the impact of the three styles of training used at NASA under common conditions. Of skill‐based training, task‐based training, and knowledge‐based training, the CREAM analysis using the new PSFs showed that skill‐based training resulted in the most significant improvement in the risk of human error, from 1 in 3 to 1 in 11. Copyright © 2013 John Wiley & Sons, Ltd.