From the Abstract: "There have been a number of high-profile incidents in recent years in which appropriate evacuation has been necessary due to the extreme nature of the incident (e.g., terrorist attacks). Evacuation under extreme situations causes specific behavioral responses and understanding these can significantly affect the efficacy of an evacuation, potentially saving lives. This paper offers a view of the current understanding of this area of research by performing a literature review into the required areas (sociological, psychological and computer modeling). It finds that although there has been significant progress in sociological models of human behavior, there is a lack of appropriate software models for extreme emergencies that can accurately model states of fear. This point is independently established in interviews with key stakeholders in the area of human behavior and emergencies. The interviewees identify that a better model is required and provide their desired areas of progress, specifically that more data is required to develop new theories and calibrate software models."