COPING WITH SUBJECTIVITY IN VULNERABILITY ASSESSMENT

Vulnerability assessment models are widely used to systematically evaluate the performance of complex safeguards systems against a variety of threats. These models require varying levels of detail and input data about the physical design of a facility and its safeguards operations and procedures. However, to evaluate safeguards effectiveness and give a performance rating, these models require additional performance data reflecting probabilities of detection, assessment, interruption, and neutralization, as well as the associated times for various adversary scenarios. These data may be attained from equipment design specifications, laboratory testing, expert judgment, or component testing. Regardless of how these data are obtained, they are inherently subjective. Although vulnerability assessment models for safeguards performance evaluation rely on many subjective inputs, their results are extremely useful in identifying strengths and weaknesses in proposed and existing safeguards systems and in evaluating the benefit of proposed safeguards upgrades. Thus, in using vulnerability assessment models it is important to ensure the quality of subjective inputs and proper interpretation of the results. This oaoer addresses the uses of various vulnerability assessment models and the nature of subjectivity in those models. The paper also describes methods for coping with subjective data.